Trachoma and Visual Impairment in the Anangu Pitjantjatjara of South Australia

L.t" )

TRACHOMA AND VISUAL IMPAIRMENT IN THE ANANGU PITJANTJATJARA OF SOUTH AUSTRALIA

A Thesis Submitted for the Degree of

Doctor of Medicine

The UniversitY of Adelaide

Department of CommunitY Medicine

Nigel Stocks B.Sc., M.B.B'S

August, 1992

¿l n ol \cl i.¡ É\r,,r o e ,'\ TABLE OF CONTENTS

Page

Abstract v

Statement vii

Acknowledgments viii

Chapter 1. Literature Review: Eye Heafth and Disease 1

1.1 lntroduct¡on 2

1.2 Assessment of Vision 3

1.2.1. Visual AotitY 4

1.2.2. Causes of Blindness in the World 5

1.2.3. Methods for Obtaining Data for Blindness 7

'1.2.4. lnternationalComparisons 10

1.3 The Major Causes of Blindness in AboriginalAustralians 12

1.3.1. Trachoma 12

2.1.1. Overview 12

2.1.2. History 13

2.1.3. TrachomaBiology 14

2.1.4. Classification of Trachoma 16

2.1.5. Acquisition of Trachoma 17

2.1.6. Trachoma as a Multisystem Disease 18

2.1.7. DescriPtiveEPidemiologY 21

2.'1.8. Transmission 31

2.1.9. Trachoma Treatment and Prevention 32

'1.3.2. Cataracts 43

1.3.3. Eye Trauma 53

1.4 Methodological Problems in Eye D¡sease Epidemiology 56

1.5 Conclusion 68 Page

Chapter 2. Blindness and Trachoma in Australia with Reference to the Anangu

Pitjantjatjara of South Australia 70

2.1 lntroduction 71

2.2 Blindness in Australia 72

2.3 Origins of Trachoma in Australia 79

2.4 Past Surveys in South Australia 82

2.5 Discussion 94

Chapter 3 A Survey of Trachoma and Blindness in the Anangu Pitiantiatjara: The

1989/90 Survey 96

3.1 lntroduction 97

3.2 Methods 98

3.2.1. Geographic and Climatic Description of The Study Area 98

3.2.2. Description of the Study Population 101

3.2.3. Description of the Collaborative Organisations 104

3.2.4. Study Design 106

3.2.5. EthicalAPProval 108

3.2.6. The Survey 108

3.2.7. Data Handling 114

3.2.8. lnformationdistribution 117

3.3 Results 118

3.3.1. Description of the SamPle 118

3.3.2. ParticipationRates 120

3.3.3. Overview 124

3.3.4. Binocular Blindness 127

3.3.5. Monocular Blindness 135 3.3.6. VisualAatity 140

3.3.7. Trachoma 142 Page

3.3.8. Service Role 153

3.4 Discussion 155 3.4.1. ParticiPationRates 155 3.4.2. VisualAotitYResults 158 3.4.3. Trachoma Results 163

Chapter 4. Changes in the Prevalence of Trachoma and Blindness in the Anangu

pitjantjatjara: A Quantitative Compadson Between the 1989/90 Survey

and Previous SurveYs 168

4.1 lnlroduction 169

4.2 Methods 169

4.3 Comparison with the 1989/90 Survey using three data sets 177

4.4. Discussion 205

Chapter 5. Research lssues 212

5.1 lntroduction 213

5.2 1989/90 SurveY lssues 214

5.2.1. Practical Problems in the Communities 2',14

5.2.2. Analysis of lntra-Observer and lnter-Observer Enor 221

5.2.3. Effect of Magnification vs Non-Magnification 225

5.3 Methodotogical Problems in the Survey and the Comparisons 230

5.4 Ethical lssues: Research in AboriginalCommunities 236

5.5 Discussion 240

Chapter 6. Conclusions and lmplications 243

6.1 lntroduct¡on 244

6.2 Discussion 245

6.3 Future Directions 249

ill Page

251 List of References

278 Appendices 279 Appendix 1. Tables of Visual Acuity 283 Appendix 2. Estimates of Death Rates 284 Appendix 3. EYe Examination Sheet 286 Appendix 4. The Five Sign Trachoma Grading System 289 Appendix 5. Ethical Approval: Letters and Permits 293 Appendix 6. MaP 294 Appendix 7. Community Newsletter: Article on Trachoma

M Abstract

This thesis examines the epidemiology of eye disease in the Anangu Pitjantjatiara people of

South Australia.

A literature review examines the methods for assessing visualdisability. The major causes of blindness in the Australian Aboriginal population and the methodological problems associated with eye health surveys are discussed.

Past surveys of blindness in Australia are reviewed, demonstrating the imperfec{ knowledge of this area. Data relating to the Aboriginalpopulation are presented and used in subsequent chapters for comparative purposes. The past causes of blindness in the Aboriginal population, with an emphasis on trachoma, are also reviewed.

The origín of trachoma in Auslralia is discussed, with the conclusion thal although it may have been present in Australia for some time, the panern of disease demonstrated by successive surveys indicates that it was not endemic in the north-west of South Australia prior to the coming of Europeans.

The design and conduct of a population-based prevalence survey is described. lt examined the prevalence of trachoma, visual impairment and blindness in a sample of 1,514 individuals, aged 0-90 years of age, representing approximately 58% of the Aboriginal population in lhe

Pitjantjatjara and Yalata lands.

Active inftammatory trachoma was found in 266/1514 (17.6V"1 ol the sample, cicatricial trachoma in 382/1514 (25.2%l and blindness (Australian definition) in 2211514 (1.5%) of the sample. The major causes of monocular and binocular blindness were found to be trachoma, calaracts and trauma. Women were found to have a increased prevalence of blindness and

severe trachoma than males. The service role of the survey is also described.

The results are compared with those of the Nat¡onal Trachoma and Eye Health Program

(N.T.E.H.P.) survey of 1976, which was performed in the same area. The comparison

v includes a discussion of the methodological problems, and a comparison of grading schemes, the sample and source populations, and the resuJts between the two surveys' A statislically significant decline since 1976 in the prevalence of binocular blindness, poor vision, inflammatory and cicatricialtrachoma was found. These results were consistent across three sets of data.

An analysis was performed of intra-observer and interobserver enor in the survey. The use of

duplicate examination sheets for this purpose is described, with the conclusion that there was

only a small amount of variation between observations for trachoma grading and visual acuity

testing. ln another study, comparing the grading of trachoma with and without magnilication,

no differences between the two methods was demonstrated.

The ethical issues of research in Aboriginal communities is discussed with relerence to the

handling and ownership of data. Several practicalproblems which arose whilst performing the

survey are documented and a review ol the methodological issues that were apparent is

undertaken. The main implications of the research findings are discussed, these include, the

treatment and control of trachoma; the provision of services for women; the evaluation of inter-

observer and intra-observer errors in surveys; the use of loupes for grading trachoma, and

aspecÍts of the survey that ensured its success in an Aboriginal community.

v¡ STATEMENT

This thesis contains no mater¡al which has been accepted for the award of any other degree or diploma in any University or College.

To the best of my knowledge and belief this thesis contains no material published or written by another person except where due reference is made in the text.

The author consents to the thesis being made available for photocopying and loan if applicable, if accepted for the award of the degree.

NigelStocks

vll ACKNOWLEDGEMENTS

I am indebted to several individuals and organisations in the completion of this thesis.

To my supervisors Dr. Janet Hiller and Dr. Henry Newland who provided praclical, field and theoretical assistance. To the South Australian AboriginalTrachoma and Eye Health Program for logistical support and to the several ophthalmolog¡sts who accompanied me on survey trips.

I would also like to thank the statf and Department of Community Medicine of the University of

Adelaide, the Department of Primary Health Care and the Department of Ophthalmology, of

Flinders Medical Centre, who gave encouragement and material support throughout. My

thanks are also extended to Lesley Hutton who helped type the thesis and to Alison Carter

and Phil Ryan who proof read this thesis.

Finally a special acknowledgement to the Anangu Pitiantiatiara, Nganampa Health Service and

yalata Health Service, without whose cooperat¡on and patience this review would not have

been possible.

vilt Chapter 1

Chapter 1. Literature Review: Eye Health and Disease

1 1.1 lntroductlon

This thesis examines the eye health of the Abodginal people (the Anangu P¡t¡ant¡at¡ara) ¡n the

north-west of South Australia. The study arose from a telephone conversation with Dr. Henry

Newland early in 1989. A statewide blindness and visual impairment prevalence survey was

being undertaken, it used a multi-stage sampling technique of people aged 55 and older who

were listed on the electoral roll. lt was felt that a concurrent survey of the Aboriginal population

was both timely and necessary, because the South Australian AbodginalTrachoma Program

wanted to assess the progress of its work and obtain new base line data lor the planning of

future work. The eye health of the Anangu Pitjaniatiara had been documented in several eye

surveys. ln the mid 1960s Mann surveyed two communities in what is now the Anangu

pitjanjatjara (AP) lands (Mann and Rountreel968). Later in the same decade, Moore et al.

(1965), examined people from urban and rural areas, and included two communities from the

north-west of South Australia. During the 1970s the National Trachoma and Eye Health

program (N.T.E.H.P.) sponsored by the Royal Australian College of Ophthalmolog¡sts

(R.A.C.O.), and financed by the Federal government, comprehensively surveyed the rural

Aboriginal population of Australia and the Anangu Pitjantjatiara were amongst the first to be

examined (R.A.C.O. Report 1980). Finally in 1985, a National Trachoma Program review

included several A P communities in a comparison with the results of the 1976'1979

N.T.E.H.P. survey (Trachoma and Eye Health Program (T.E.H.P.) Report 1985).

These surveys, particularly the last two, demonstrated the poor eye health of Aboriginal

people in Australia. The surveys also identified three common eye diseases which wers

responsible for the high prevalence of blindness found: trachoma, cataracls, and eye trauma'

Although several cross-sectional surveys have been undertaken, only one comparative study

(the 1985 review), had been attempted previously in South Australia, and it was limited to an

assessment of trachoma. There has been one other comparative study in Australia (in the

Northern Territory), it examined changes in trachoma from the 1940s to the 1980s (Meredith

et al. 1989). No surveys have assessed visual acuity changes over lime and although

trachoma is a marker of eye health and socioeconomic development (section 1.3.1.), and

2 Chapter 1 could be used solely to assess changes in the eye health of a population, an analysis of changes in visual impairment, adds to any comparative description.

The following literature review examines several areas of relevancs to the assessment of eye health. The assessment of visual acuity, methods of obtaining data on blindness and the difficulties with internat¡onal comparisons of eye disease or blindness prevalence. The biology and classification of trachoma is reviewed to establish background information on its epidemiology. These include elements of particular relevance to the atfected Aboriginal population; the way it is acquired, its ability to infect multiple sites in humans, its descriptive epidemiology, transmission, and methods of lreatment or prevention (section 1.3). The comprehensive picture of trachomatous infection established is used in chapter six to discuss the control and possible eradication of trachoma from rural Abodginal communities.

Another element of the literature review concerns the methodological problems in eye disease epidemiology. The problems associated with cross-sectional prevalence surveys is discussed, and the ditficulties in using this type of study for comparisons over time examined.

1.2 Assessment of vlslon

ln the examination of the eye an assessment of v¡sual acuity is essential, both as an aid to

clinical diagnosis and for following the progress of an eye disorder'

ln eye surveys an assessment of visual acuity can be the lirst step ¡n screening the populat¡on

for blindness, or for ocular condílions such as catarac'|. Once identified, lhe atfected people

may then be examined, a cause diagnosed and treatment with follow-up ananged. lt is also

important to identify ditferences in the delinition of blindness, and to recognise that the ma¡or

causes of blindness, which reflect genetic, socioeconomic, climatic and ophthalmic care

factors, vary between countries. Additionatly in eye epidemiology, it is important to be aware

of problems with blindness estimales when making internationalcomparisons.

3 Chapter 1

1.2.1. Vlsual AcultY

The ability to see is an important sensory function because vision shapes our perceptions of the world and atfects our interaction with it. Good visual acu¡ty is centralto survivalin hunter- gatherer societies, and blindness in developed countries is a feared disability. Previous surveys have documented the very good visual acuity of Australian Aboriginal people, particularly in the younger age-groups (Edrruards et al. 1976). However, older Aborigínals have rates of blindness comparable to those found in developing countries (R'A.C.O. Repod l9g0). lt is a tragedy, therefore, that a signilicant proport¡on of older Aboriginals from rural areas are condemned to years of poor vision or blindness.

Although the standard for testing visual acuity is well defined and comparable internationally, blindness has many definitions. These definitions vary from country to country, and may depend on the residual visual acuity in the better eye with correction, or the degree of impairment in the field of vision. Some countries rely on a func{ional definition rather lhan a

measure of sensory capacity (Goldstein 1980). The World Health Organisation (W.H.O') delines blindness as "a corrected visual acuity in the better eye of counl fingers at one metre

or worse. (W.H.O. 1973). The comparable current Australian definition is less reslrictive: 'a

corrected visual acuity in the better eye of less than 6/60" (SocialSecurity Department 1991).

Differences in objective definitions, and the reliance by some countries on flexible funclional

assessments, reflec{ economic, social and technological factors. The cost of providing

blindness pensions or facilities allows only Western countries a liberal ¡nterpretation of

blindness. Socio-cultural factors emphasizíng family support may be more evident in Th¡rd

World countries which lessens the need for governmental support to those with visual

disability (Goldstein 1 980).

4 Chapter 1

1.2.2. Causes of Bllndness ln the World

It has been estimated that there are at least 28 million blind people in the world (Thylefors et at.1987). The majority (90%) live in Third World counlñes where blindness rates are 10 to 20 times higher than in developed nations. Using epidemiological models, a blindness rate of

.l7o occurs in developing countries with endemic trachoma, foci of xerophthalmia or

onchocerciasis and insufficient eye care facilities. A rate of 0.5% has been reported in

countries of intermediate development where there are problems with the provision of eye

care services, particularly cataract surgery. ln developed countries a rate of 0.2% is

considered representative (Thylefors 1990).

Blindness is most often found in rural areas where socioeconomic conditions are poor,

medicalfacilities meagre or non-existent, and basic services (e.9. sewerage, power) lacking.

Over two-thirds of existing blindness is avoidable, either correctable with cataract surgery, or

preventable as with trachoma (Sommer 1988)'

World-wide, the major causes of blindness are (Thylelors 1987)

Trachoma (section 1.3.1.)

Trachoma infection commences in infancy in endemic areas, with reinfection and lack of

treatment teading to progressive scarring of the conjunctiva. The eyelashes invert (trichiasis),

louching the cornea, causing corneal opacity and eventual blindness.

The intensity of the inllammation in its act¡ve stage, as well as its prevalence, is important in

predicting the risk of blindness. Socio-cultural, as well as economic factors atlect its

prevalence in a community. Whilst economic and simple hygiene changes have been noted

to reduce its impac{,lreatment campaigns us¡ng tetracycline eye drops, system¡c antibiotics,

and surgery for trichiasis, are needed in areas where there is a high prevalence of trachoma.

5 Chapter 1

Xerophthalmia

Vitamin A deficiency affecls children up to five years of age. Because mortality is high in

severe blinding disease, its impact may be underestimated. Seasonal and geographic

variations occur owing to the variable availability of food containing vitamin A and to dietary

habits in regional Populat¡ons.

Cataract (section 1 .3.2.)

Cataract is recognised as the mosl common cause of visual loss in the world. lt is closely linked

to ageing. Whilst in Western countries it is readily correcled with surgery, in developing

nations an already established backlog and the lack of medical facilities combine to make

cataracts a major public health problem. Risk lactors for cataract have not been conclusively

identified. Preventive measures or primary interventions are speculative, and therefore

surgery remains the onlY oPtion.

Onchocerciasis

Onchocerciasis is a parasitic disease which directly atfects the eye. lt is limited to savanna and

forest locations in Africa and some areas of Central and South America. Chemical controlof

vectors is needed. Treatment requires ophthalmologicalreview and chemotherapy.

Glauæma

Glaucoma presents in the majority of cases in one of two forms; an acute, angle closure

glaucoma with symptoms of pain and loss of vision, or in a chronic form in which the person

has a progressive loss of visualfields. lt is related to increasing age and is a growing problem

in Western countries.

Ocular trauma (section 1.3.3.)

Trauma is otten a cause of unilateral loss of vision. ln developed nations it is associated with

industrial and motor vehicle accidents. ln Third World countries minor ocular trauma due to

6 Chapter 1 foreign bodies may, il neglected, cause secondary bacterial infection and lead to corneal

scarring and blindness in the affected eye.

Refraclory anomalies and amblyopia

Severe myopia may functionally impair a person when suitable corrective lenses are not

available. Non-conection can lead to a blindness in the affected eye.

1.2.3. Methods for Obtalnlng Data lor Bllndness

Sources of information on the prevalence and causes of blindness vary between countries.

The theorelical methods include (Goldstein 1980):

Censuses

Although every household is approached, thus theoretically covering the whole population,

problems exist with:

1. Self or proxy reporting of information.

2. Objective definitions of blindness.

3. Possible exclusion of institutionally blind, chronically or mentally ill people

4. lnadequate or lack of recording of the cause of blindness.

5. No provision for assessing visual acuity'

It might be possible, in countries such as Australia, to obtain data on the number of people

receiving blindness pensions, but ditficulty has been encountered with data when the legal

de¡nitíon of blindness was not used consistently and no clinical examination carried out

(Banks and Hutton 1985). Basic socio-demographic data could; however, be obtained,

enabling better assessment of needs.

7 Chapter 1

Estimates

Relatively inexpensive estimates are based on previously collected data with extrapolation, either over time, f rom one year or decade to the next, or over geography, from one counlry to another or one area to another within a country.

Extrapolation could also occur from one data source, say prevalence of diabetes, to another, such as the expected prevalence of diabetic retinopathy.

Several assumptions often need to be made regarding age, sex and racial distributions which can reduce the usefulness of the information. ln addition, the known or unknown accuracy ol the original source may make the accuracy of conclusions reached impossible to estimate.

Registers

Blindness Registers have several advantages (Goldstein 1980):

1 . They are sources of acæurate information on age, Sex, race and Cause.

Z. They enable trends to be detected as continuous reporling is involved.

3 Wth continual updating (eliminating those who have died, rnoved on, or had

their vision restored), acc¡rate prevalence data may be est¡mated.

4. Service provision may be based on register data.

There are a number of disadvantages, which include:

1 Self selection of registrants may affect the applicability of the statistics to lhe

general community.

2 Voluntary registration implies there must either be an advantage to the blind

person, the referring ophthalmologist or optomelrist. This ls linked to the

availability of services, which may be subject to variation over time, and may be

applicable lo certain individuals only e.g. socially disadvantaged groups.

I Chapter 1

3 Newly registered blind may rþt be recerìtly blind, and may reflect a delay,

either by the person or ophthalmologist, in recordíng the information,

therefore not providing uselul incidence data.

4 Registrants may differ between geographic areas, making cross-comparison

difficult.

5 The function of the Register is to provide services and not data. Unless a

specific effort is made to ensure the completeness and accuracy of the

information, its use for statisticalpurposes may be invalid.

Surveys

A census survey, whilst advantageous in terms of population coverage, is impractical with

large numbers of people. Sample surveys, using a random subset of individuals who are

representative of the ent¡re population, are more economical and theoretically just as

accurate.

Prevalence rates may be estimated from sample surveys, and if clinicalinformation is obtained

by examination, then accurate data on visual acuity and causes of blindness may be obtained.

However, incidence data are not obtainable unless surveys are repeated at short intervals.

Also institutional populations are often ignored and any degree of self selection, including

refusal to participate, can bias the result.

The W.H.O. has developed survey methodology which incorporates several elements

(Thylefors 1987):

1 . A preliminary local census, usually on a house to house basis.

Z. Preliminary screening of visualaority at the 6/18 level, in accordance with the

internationally accepted definition of low vision. Further testing is performed

9 Chapter 1

if the threshold is not passed, to establish the level of visual aotity (the

Snellen "illiterale E" chart is used).

3. A basic eye examination using simple x2 loupes.

4. Grading trachoma or xerophthalmia if appropriate for the country.

S. A detailed eye examination on those individuals who are blind or have low

vision.

6. Assessment of the need for treatment or further care

1.2.4. lnternat¡onal Comparlsons

Demonstration of stat¡st¡cal ditferences between countries in the prevalence or causes ol btindness has its probtems. Although great care may have been taken in collecting the data from registers or conducting surveys, variation may be attributable to ditferences in definition, classification of disease, or diagnostic acumen, or to a range of sampling problems. Goldstein oulined several areas of importance (Goldstein 1980). Problems occur due to ditferences in definition, in coverage and to differences ¡n ophthalmological resources. There are also problems due to ditferences in classification and methods of data collection.

A careful assessment of both the results and the methodology is therefore required for valid internationalcomparisons to be made. These comparisons can be useful in determining the aetiology of an eye disease when it is differentially distributed between countries.

Summary

Vision is an important sensory function. Visual acuity test¡ng has been internationally

standardlzed. Although the W.H.O. definition of blindness is well known, individualcountries

10 Chapter 1 adopt standards that are applicable to the socioeconomic conditions prevailing in their regions.

The major causes of blindness throughout the world vary, but most are preventable or readily treatable, indicating that there is a great burden of unnecessary blindness in the world.

There are several methods for obtaining information about the prevalence of blindness, each has methodological problems which either must be overcome or acknowledged in presenting the data. lnternational comparisons must therefore be carefully scrutinized to ensure tho validity of the conclusions.

11 Chapter 1

1.3 The Major Causes of Bllndness ln Aborlglnal Australlans

1.3.1. Trachoma

Trachoma has been previously identified as an important cause of visual disability in the

Australian Aboriginal population (R.A.C.O. Report 1980). lts relationship with socioeconomic development, social and cultural factors has been demonstrated (Man 1989). An

understanding of its history, biology, and epidemiology is therefore important if treatment and

prevent¡on programs are to be etfective.

2.1.1. Overvlew

Trachoma is a term that describes both the acute infection caused by the oqanism Chlamydia

trachomatis, and the long term sequelae of repeated infection which are, trichiasis or inversion

of the eye lashes, and corneal opacity.

Trachoma affects approximately 500 million people world wide (M.M.W.R. 1982), mainly in the

rural areas of the tropical and sub-tropical zones. Trachoma is associated with adverse living

conditions and poor socioeconomic development. Although readily treatable with antibiotics

the disease remains endemic in many underdeveloped nations (West and Taylor 1988,

Carmichael et al. 1982). Six to nine million people are estimated to be blind from trachoma,

with many more sutfedng partial loss of vision (Dawson et al. 1981 a ). lt is the leading cause of

preventable blindness in the world today (Tabbara 1990).

A review of current literature on trachoma suggests that without any immediate prospect of a

vaccine, or of rapid socioeconomic development, lts effeclive eradication will rely on

community-based action, with the role of the water supply, sewage disposal, and traditions of,

and facilities for, personal hygiene being ol equal or greater ¡mportance than the use of

antibiotics (Smith 1991 ).

12 Chapter 1

2.1.2. Hlstory

Trichiasis was known and treated in China in the 27th century B.C. and trachoma was thought to be endemic among the Sumerians of Mesopotamia before 2,000 B.C. (Duke-Elder 1965).

Trachoma was demonstrated by Ebers in 1889 in evidence from papyri, to be present in Egypt

3,500 years ago (MacCallan 1913). Hippocrates included some works on trachoma and

trichiasis in the Corpus Hippocraticium (Mettler 19471. An Alexandrian ophthalmologist,

Heliodorus, in the 2nd century 8.C., described a surgical procedure for trichiasis (MacCallan

1931). ln Rome, Celsus wrote about trachoma "aspiritudo" at the time of emperor Nero. He

maintained the Greek notion of trachoma as a disease causing runny eyes with inllammalion

and granutation of the inside of the eyelids (Mettler 19471. lt appears that l¡ttle progress

occurred ¡n ophthalmology from the 2nd to the 7th centuries; however, from the 8th to the

14th centuries A.D. there were at least 60 Arabic ophthalmologists who described and treated

trachoma. Very little was documented ¡n the years between the decl¡ne of Arabic science and

the emergence of trachoma in Europe in the 1900s (AÞRifai 1988).

Trachoma was probably introduced into Europe for the first time by crusaders returning from

palestine (Darouger and Jones 1983), ¡t was reintroduced after the Napoleonic Wars in Egypt

(1798-tBO2), being spread quickly by the returning French and English troops (Maccallan

191g), and was responsible for repeated epidemics in the early decades of the last century

(Botdt 1904). Moorfields Eye Hospitalwas established in 1805 to treat the trachoma found in

the East End of London (Jones 1980).

There was much dispute in scientific circtes about its aetiology and transmission. Some held it

to be non-contagious and defended the theory of a spontaneous miasmic origin. However

with advances in bacteriology and other scientific knowledge, the theory that trachoma was an

infectious disease spread by close contact was gradually recognised (Boldt 1904).

However primitive such theories seem in retrospect, some of the observations and

experiments performed ¡n the late 1gth century were remarkably accurate. Researchers and

clinicians identified family, socioeconomic and personal factors that were important ¡n the

13 Chapter 1 epidemiotogy of trachoma (Boldt 1904). Specific suggestions were made for improving in hygiene among the troops under Napoleon (Feibel 1983), and among the people who lived the tropics (Elt¡ot 1920). This indicates that although the disease was not well understood, some principles of its prevention were.

2.1.3. Trachoma BlologY

Trachoma is an infectious disease caused by the organism Chlamydia trachomatis. Chlamydia

RNA, and which is an obligate intracellular bacterium which possesses a cellwall both DNA and which, divides by binary fission. The organism has two alternating forms: the elementary body although infective, is not metabolically active or capable of replication, and the initial body' which cannot survive outside the host cell, but can divide and is metabolically active' The organism was first grown by Tang et al. in Peking (1954.

lnfection with chtamydia trachomatis has been described as occurring in two patterns

reminiscent of treponemal disease, yaws and syphilis respect¡vely (Schachter and Dawson

t9g1). lt appears to be a pathogen specific for humans (Fletcher and Gordon 1990) and 15

serotypes have been identified (Schachter 1978). D-K, L, L2, L3 predominate in sexually

transmitted disease and serotypes A, B, Ba, C are invariably associated with classic trachoma

in endemic areas of developing countries (Schachter and Dawson 1981).

paratrachoma (also called inclusion conjunctivitis or inclusion blennorrhea) has been used to

describe the cross-over of sexually transmitted serotypes to ocular sites (Viswalingan et al.

19g3), which usually occur in neonates. The signs and symptoms are typical of classical

infection, and although not common this has important ¡mpl¡cations in the pathogenesis and

treatment of chlamydia inf ections.

14 Chapter 1

The clinical features of ocular inlection are as follow (Darougar and Jones 1983) lncubation Period

The incubation period is one to three weeks as demonstrated in human volunteers.

Symptoms

The usual features are: eye irritation, a foreign body sensation, itching, and a watering mucopurulent discharge with redness and swelling of the eyelids. From the experience of ophthalmologists in the field, it appears that rural populations otten accept the symptoms of inf ection without comPlaint.

Signs

The signs have been described (Tabbarra 1990) as: periocular lymphadenopathy which rarely

occurs, except in the early form, and a supelicial pedpheral and supedor keratitis followed by

pannus formation (Pannus is a superficial vascularization of the cornea with infiltration ol

granulation tissue). Limbalfollicles can occur, which healto lorm charac{erist¡c Herberts pits,

which are scars of the semilunar area of the limbus and appear as small punched out holes.

Follicles occur which may be pinhead, localized or whitish-grey elevated nodules. They occur

on the conjunctiva and sometimes there is ditfuse oedema and a papillary reaction associated

with a conjunctivitis.

The natural history has been briefly described as an initial inflammatory stage ¡n early

childhood with dilfuse papillae involving the conjunctiva, usually seen on the upper tarsal

plate. Repeated infection, followed by chronic inflammation, then leads to conjunctivalscars,

contracture of the lids with inversion of the eyelid (entropion) and ingrowing (tdchiasis) of the

eye lashes. This eventually progresses to corneal scarring, corneal opacity and visual

impairment. Secondary bacterial involvement can hasten the process. Blindness can be the

final result if the trachomatous infection is lett untreated.

15 ChaPter 1

2.1.4. Classlflcatlon of Trachoma

One of the first trachoma classifications developed was based on Macoallan's several years of experience working in Egypt (MacOallan 1913).

Trachoma I - seen typically after Chlamydia trachomatis infection has

occuned. Greyish follicles are seen on the upper tarsal plate.

Trachoma ila - in which follicles predominate, otten associated with ocular

discharge.

Trachoma ttb - in which papillary hypertrophy coexists with the lollicles

Trachonn ttc - trachoma complicated by chronic gonococcalconjunctivitis

Tracho¡v ttt - in which cicatrization (conjunctival scarring and contraclure of the

upper tarsal Plate) is beginning.

Tnchoma lV - in which cicatdzation is complete.

Several other classifications have been developed (Vedn et at. 1987); however, they have

often been complicated and suited only to specialist interpretation . Although detailed, the

reliabitity of these ctassifications has been questioned (Btilliant et al. 1983 b ), and modified

versions have been tried (Telsch et al. 1987). Recently a simplified five sign system has been

developed to overcome the shortfalls that were evident in previous systems (Thylefors 1987,

Taylor et al. 1987). The new classification, as its name implies, has only five signs that need to

be recognized as being present or absent. These are trachoma follicles (TF), trachoma

¡ntense (Tt), trachoma scarring (TS), trachoma trichiasis (TD and cornealopacity (CO). The

advantages of the five sÍgn system are: it avoids grading the severity of a particular sign and

therefore is easier to use; and in inter-observer trials it was shown lo reduce interobserver

errors (Taylor 1987 b ). The grading system is outlined in appendix 4.

16 Chapter 1

2.'1.5. Acqulsltlon ol Trachoma

The early acquisition of trachoma is inlerred from prevalence studies which show infection soon after blirth, with peaks at 1-2 years of age (Dawson et al. 1976). Scaning has been shown to occur in 3-6 year olds, with 50% showing scarring at age 10 in endemic areas (Kupka et al.

1968). However in Tanzania, little scarring is observed in children and the peak prevalence of active disease is among 3-5 year olds (West et al. 1991). The acquisition of trachoma can therefore be seen to vary between endemic areas.

The progression of trachoma is that of a chronic infection with intermissions and relapses.

Observations in animal experiments have suggested that periodic reinfeclion plays a role in producing the cicatrization reactions that lead to corneal scarring and lid distortion

(Monnickerdam et al. 1980 a+b and Monnickerdam and Pearce 1983). A longitudinal study of trachoma incidence ¡n three groups of school children in Taiwan showed many examples of

intermittent trachoma infection, apparently due to reinfection or relapse (Woolridge et al.

1967). More recently, reinfection has been viewed as the main cause for relapse. ln a len

year study, again in Taiwan, 32 households containing adults and children, were followed with

repeated clinical and laboratory observat¡ons, lhe study demonstrated that progression of

trachoma to pannus and scaning occuned only atter reinfection (Grayston et al. 1985).

Epidemiological studies in South Africa, looking at household clustering and serological

evidence of lrachoma infection, support the idea of a multicyclic disease based on the family

unit, that is, a dísease that recurs in a group of people living together (Ballard et al. 1983). ln

lran, Treharne showed that re¡nfection within households was by lhe same Chlamydial

serotype, with the bulk of infective organisms being shed from the eyes of infants and young

children (Treharne 1 985).

The multicyclic nature of trachoma implies not only reinfection but spontaneous healing as an

¡mportant aspect (Grayston et al. 1985). Researchers followed 877 children over several years

and demonstrated reversion to normal eyes in one third of lhose with trachoma at entry to the

study, and of the 43% who developed trachoma during the 6 years study, again one third

17 Chapter 1 reverted to normal (Woolridge et al. 1967). They also dernonstrated, in pñmary cases, that the clinical disease diagnosed as lrachoma lasted from less than one to two years (average one yea4. ln all but one case, tear antibodies disappeared close to the time classical disease ceased, but serum antibodies persisted longer. ln no single case did clinicaldisease progress beyond trachoma stage l. ln secondary cases in older age-groups, all showed signs of corneal vascularization, but when it became inactive five of nine showed no signs of trachoma' ln tertiary infect¡ons in those cases with inactive stage lV trachoma, tho course of infec{ion was brief but lasted in some as long as secondary ¡nfection. There was evidence of disease transmission amongst family members in eight of the tertiary infections.

The occurrence of secondary bacterial infection of trachomatis eyes has been accepted for many years. Reports ¡ndicate that secondary bacterialinfections may play a pad in determinirq the outcome of long-standing trachoma (Jones 1980). Cerlainly untreated bacterial inleclion can cause visual impairment and blindness (Reinhands et al. 1968) and an increased rate of recovery of bacleria has been demonstrated from the lids and conjunctiva of patients with moderate trachoma (Chumbley and Thompson 1988)'

2.1.6. Trachoma as a Multlsystem Dlsease

Among adults Chlamydia trachomatis has been reported to cause cervicitis, salpingitis,

endometritis, vaginitis, peritonilis, Reiter's syndrome, arthritis, endocarditis, epididymitis,

urethral syndrome and urethritis (Fletcher and Gordon 1990).

Speo.rlation thal Chtamydia trachomatis might not only be ¡solated from the eyes, but may also

be found at other sites, occurred in the 1950s. Abu-Jaudeh working in Lebanon, suggested

that trachoma might be a generalised disease. This was supported by evidence showing

inclusion bodies in other mucous membranes (Abu-Jaudeh 1953). ln Australia, during several

trachoma surveys ¡n the Northern Territory, Flynn noted a chronic nasaldischarge in most of

the Aboriginalchildren with trachoma, and observed that systemic therapy was rnore effective

18 Chapter 1 than topical therapy alone (Flynn 1957). He suggested from these observations that trachoma might be a generalized disease.

However, direct evidence that Chlamydia trachomatis could be cultured from extraocular sites in children f rom trachoma endemic areas was not available untilthe 1980s. ln Egypt, a survey

attempted to isolate Chlamydia trachomatis from the eyes, nasopharynx and rectum of 128

chitdren from three villages (Malaty et al. 1983). One third of all children yielded Chlamydia

trachomat¡s on a single sample. Of these 43 children, twelve (28o/"1with conjunctival infection

had Chlamydia trachomatis elsewhere, and seven (16%) had negative eye cullures but had

either nasopharyngeal or rectal infection.

Transmission oÍ Chlamydia trachomafis from mother to child during childbirth was first

demonstrated by Jones et al.(1959), Chlamydia trachomatis was cultured from both eyes of an

infant with inclusion conjunctivitis, and from the cervix of the mother.

The importance of perinatal transmission with respect to more generalized infection was

demonstrated in 1975. A case oÍ Chlamydia trachomatis pneumonitis was reported in a child

who had been treated for inclusion conjunctivitis of the newborn (lCN) (Schachter et al. 1975).

Subsequent studies have demonstrated lhal Chlamydia trachomatis is the commonest cause

of neonatal conjunclivitis (Rapoza et al. 1986), and has also been associated with acute lower

respiratory tract infection in children 3-4 months of age (Paisley et a|.1984). ln a recent study

of 3309 births in which 198 mothers had a chlamydia infection, ol 174 mothers and infants

folfowed, 43 infants were positive lor Chlamydia trachomat¡s. A total ol 24 had symptomatic

inlection and 19 asymptomatic chlamydia infection. A respiratory infection occurred in 6

(14y"1, conjunctivitis in 20 (46.5%), and otitis media in one. The overall incidence of

Chlamydia trachomatis infection was calculated as 8.2/1000 live births (Preece et al. 1989).

Similar results from another prospective study in America has lead to the suggestion that

screening lor Chlamydia trachomatis in pregnancy should be routine (Schachter et al. 1986 a

). ln trachoma endemic areas the piclure may be different. A study in Kenya, demonstrated

that the trachoma serovars isolated from the genital tract of mothers and the eyes of their

19 Chapter 1 children, were different, and the authors concluded that perinalaltransmission was not a major route of transmission. However, the validity of the study was limited by the difliculty in culturing Chtamydia trcchomatis from the eyes of children, and the high number of mothers who declined speculum examination (Brunham et al. 1990).

Chtamydia trachomatis has also been implicated ¡n otitis media in children. Although it is a comrnon complication in infants wilh Chlamydia trachomatis pneumonia, and headng loss has been associated with otitis media in adults, the contribution of Chlamydia trachomatis to otit¡s media in children 4 to 5 years of age is debatable (Schachter and Dawson 1981). ln Australia,

The Menzies School of Health Research, using evidence obtained from a prevalence survey ol Chtamydia trachomafis conducted in two Aboriginal communities, found that ear perforations occurred at an earlier age in communities with: more trachoma in infants, early chlamydial infection, less passive antibody against chlamydia and less usage ol antibiotics

(Douglas et al. 1987 (19) ). The authors concluded that the data implicated chlamydia as a cause of chronic otitis media; this was despite the facl that they isolated Chlamydia trachomatis in only one of 119 swabs. They remained convinced of its causal role for the following reasonst it was olten grown from the nasopharynx; other studies of middle ear aspirates had been of older children after any initia! damage by Chlamydia trachomatis infection had already occurred, culture of middle ear discharge were usually positive lor other bac{erial pathogens which may have prevented isolation ol Chlamydia trachomat¡s; and aspirates of middle ear discharge were otten positive for anti-chlamydial antibody. A second longitudinal study of

Aboriginal children found the risk of ear perforation was lowest in infants bom to rnothers with the highest levels of anti-chlamydial antibody in the maternalserum. The inference made was that the antibody was protective againsl the subsequent development of chronic otitis media, supporting lhe hypothesis that Chlamydia trachomatis is important in the aetiology ol chronic otitis media (Douglas et al. 1987 (20)).

Another study in Central Australia suggested lhal Chlamydia trachomatis might be implicated in chronic otítis media. lnstead of using culture techniques, a direct fluorescent labelled

monoclonal antibody test was applied to samples of smears taken from the eyes, throat and

20 Chapter 1 middle ears of 18 children (racially European), aged 1-11 years. The results dernonstrated a high carriage rate in all sites even though clinical trachoma was not evident. The author could not say whether the resutts indicated causat¡on or merely commensal cardage (Banks et a1.1985).

Although infection wilh Chtamydia trachomatis can occur at many extraocular sites, ils importance with respect to the lransmission and pathogenesis of trachoma is uncertain. Auto- inoculation from extraocular sites in an individual has been proposed (Schachter and Dawson

1991). lf this form of transmission can be demonstrated, it has important implications for treatment and control. lt explains treatment failures when only eye drops are used, with the reservoir of infection spilling back to the eyes lrom extraocular sources from either the infected individual or others. lt further ¡nd¡cates that systemic therapy may be more appropriate in the younger age-groups, to control not only trachoma but also the respiratory,

middte ear and possible diarrhoealdiseases which cause signilicant morbidity and mortality in trachoma endemic areas (Schachter and Dawson 1981).

2.1.7. Descrlptlve Epldemlology

The importance of Chtamydia trachomatis as an organism causing infectious disease is well

recognised. lts distribution and prevalence, despite mass treatment programs, indicates its

success as an organism. lts ability to colonize muttiple sites and remain asymptomatic for long

periods, indicates its adaptab¡lity and potential as a human parasite. Yet, like smallpox, it

appears to have no animal reservoir of importance (Schachter 1985). Eliminating it cornpletely

is possible, at least theoretically, if not on a world scale then on the community or regional

level.

Trachoma was a major cause ol blindness in Europe and North America during the earty part of

the 1gth cenlury. As mentioned previously it is a persisting problem in developing countries.

Its declirre in Europe preceded the advent of antibiotics (Rienhards 1967). As with many

21 Chapter 1 other infecfious diseases, socioeconomic development, changes in living standards and improvements in community hygiene significantly decreased the severity and transmission of trachoma.

The interaction of the host, environment, and organism has been outlined (Jones 1975).

Jones divided trachoma into three classes of infection. The first class was blinding trachoma in hyperendemic areas, with bacterial infections playing an important role in its pathogenesis.

Serotypes A, B, Ba, C predominated. The second group was defined as non-blindin! trachoma occurring in hyperendemic or hypoendemic areas. Serotypes A, B, Ba, C, and sometimes D, E and C were usually isolated from infecled individuals. Superadded bacterial

infection was not a feature of class two disease. The third classification was called

'paratrachOma'and Occurred in hypOendemic areaS. SerOtypeS D, E, P, G, H, l, J, K and

sometimes B and C were allfound in these areas'

The division of trachoma into classes has practical applications. ln the Northern Territory

differences ¡n the intensity of trachoma between the "Top End" and Central Australia were

noted as early as the 1940s (Schneider 1946). The disease was lound to be mild in the Norlh,

with complications rare, but severe ¡n the South. Schneider attñbuted this variation to climatic differences. The findings were repeated in 1957 when bacteriologic and virologic

investigations were undertaken ¡n the "Top End. of the Northem Territory. Practically no

cicatrization or pannus was found (CroW et al. 1959). ln a more recent evaluation (Meredith et

al. i989), this initial difference in severity with latitude, was reflected in prevalence surveys

over the past 40 years, with a significant decline occurring in the norlh, and only minor

changes in the south. However, the attribution of these changes solely to climate would

ignore the role of other factors in the epidemiology of trachoma.

A Wortd Health Organisation expert committes (W.H.O. 1962) listed the following possible

factors in the epidemiology of trachoma: race, climate, insect vectors, diet and nulrit¡on,

cultural and social customs, poputation density, general economic development, previous

population movements, educational status and the presence of other ocular and general

22 Chapter 1 diseases. This list has been reviewed and it was concluded that although it was comprehensive '... it did not materially advance our understanding or foster the development of specific intervenlion stralegies ..." (Taylor and Sommer 1985). However, a review of some of these factors as follows, allows an appreciation of trachoma's epidemiology.

Distribution

Trachoma is distributed in areas of North Afñca extending down to East and South Africa, into the Middle East, and into some parts ol lndia and south-east Asia. Pockets of trachoma exist in South America, Australia and the Pacific lslands. lt is a problem mainly in lhe rural areas

(Treharne 1985).

Host Determ¡nants

1. Age

Children are viewed as the main reservoir of infection (Treharne 1985), which has

been shown to occur in young babies 2-3 months of age (Dawson et al. 1976). Active

inflammatory disease in endemic areas is most commonly found in children under ten

years of age. ln a survey of a trachomatous region the cumulative incidence of

acquired infection was 73"/o in pre-school children, and 2.2"/" in the rest of the

population over a 4-year period (Treharne 1985). ln endemic areas it appears that

active trachoma predominates in the younger age-groups, and cicatrization in older

individuals.

2. Sex

ln young children there appears to be no sex difference ¡n the rate of active trachoma

(Graham et al. 1973, Jones 1975), although Hollows found a slight predominance of

folliculartrachoma in males aged 19 years or less in AbodginalAustralians (R.A.C'O.

Report f 9BO). ln older children and young adults; however, trachoma begins to

predominate in girls (Ballard et al. 1983). Severe cicatrization is more common in older

23 Chapter 1 women than men (R.A.C.O. Report 1980 and Telsch et al. 1988) and women consequenily suffer a greater loss of vision (Brilliant et al. 1985). The following environmental and behavioural factors are bel¡eved to be responsible lor this difference (KuPka et al. 1968):

- Female infants receive less care than males, resulting in increased

exPosure to infection.

- Young girls are often given charge ol younger brothers and s¡sters,

and are therefore exposed to familial sources of infection.

- The mother has closer and rnore frequent contact with younger

children than the father, increasing the chances of transmission.

The increased risk in women may therefore be due to repeated re¡nfect¡on in early adulthood, the reinfection occurring from young children. However a study examining the risk factors for trich¡as¡s in women conducted in Tanzania (Turner et al.

lggg), found that prolonged exposure to child care as a young girlor molher showed

no significant difference between cases and controls.

3. Race

Conflicting evidence has been presented on the significance of race and

suscept¡bility to trachoma lnfection. Some early authors asserted that there were rìo

ditferences among the races, with all suffering equally under similar conditions

(Macçailan 1931). Others have found hypersusceptibility among certain ethnic

groups (Guerra 1957). More recent studies in hyperendemic regions have not

demonslrated differential predisposition to trachoma between racialgroups (Trehame

1 e85).

24 Chapter 1

4. Cultural Factors

Although individuals catch trachoma, they exist within a family, which is part of a surrounding community. Cuttural pract¡ces can have an immediate effect on both the transmission and development of trachoma, also, cultural dislocation is common in trachomatous areas (Hollows 1985). Subtle variations in cultural practices, for example, atfecl the utilizatíon of water, and the prevalence of trachoma (West et al. lggg). ln a prevalence survey of lrachoma on the West Bank and Gaza Strip, there was a substantialvariation in the prevalence ol trachoma in communities which were similar geographically and socioeconomically. The authors thought many factors such as social and hygiene practices, fly populations, water supply and the availability of medicines were very important (Chumbley and Thompson 1988).

Although personal hygiene pracl¡ces have long been held to be important in the transmission of trachoma (Boldt 1904), lhese assessments were qualitative. The provision of washing facilities and daily baths was found to reduce the prevalence of trachomatous disease in an Aboriginal population in South Australia (Hardy et al.

1967). ln another study, 1097 people in two Mexican communities were surveyed lnformation was gathered on personal and lamily hygiene practices. Twenty{ive percent of children under the age of ten had significant inflammatory trachoma, and almost 100% of those older than 40 had cicatricial disease. Risk lactor analysis by contingency tables and Chi-squared testing showed that a face washing frequency of greater than 7 times per week was assoc¡ated with lower rates of trachoma compared to less frequenl face-washing (Taylor et al. 1985). Taylor concluded that the

evaluation of many different community, family, and personal practices and

behaviours could be done, and interventions designed for implementat¡on at the

community level.

Direct application of these practices to other cultural settings may not have the same

result, as demonstrated in a study from the Northern Territory. Face washing in

25 Chapter t

Aboriginal children in a remote settlement did not have any effect on the prevalence

of trachoma (Peach et al. 1987).

Additional recent work has locussed on nasal discharge and flies rather than facial

cleanliness (as measured by dust, food on face, or 'sleep' ¡n the eye), with

suggestions that targeting efforts at cleaning nasal discharge and keeping flies off

chitdren's faces would produce the biggest effect on trachoma prevalence (West et al.

1991).

Another lactor related to hygiene practices was assessed in a survey of an Egyptian hamlet.

The absence of a pit latrine was predictive of increased trachoma prevalence. lt was conctuded that the construction of pit latrines, presumably by reducing the density ol the fly population, could be a simple and acceptable method of reducing trachoma prevalerrce in some communities (Courtñght et al. 1991).

Environmental

There is contradictory evidence that climate has an association with trachoma. Although

trachoma is found today mainly in drier climates (West et al. 1989), 100 years ago it was

common ¡n the wetter areas of Europe, especially wherever conditions of poverty or crowding

existed (Reinhards 1967). Mild trachomatous disease in Jamaica was attdbuted to its well

watered environment (Dawson and Schachter 1967). However, seasonal differences have

not been studied extensivelY.

ln Australia; however, a strong association between follicular trachoma, humidity, evaporation

rate, and bright sunshine hours was found. A weaker association was found with rainfall, ultra-

violet exposure and latitude. Cicatricialtrachoma was associated with low humidity, increasing

evaporat¡on rates, bright sunshine hours and rainfall (R.A.C.O. Report 1980). Control for

conlounding factors such as water quality was not done. The availability of water was linked to

both these conditions, and also to poor sanitation, increased flies and dust. Water quality may

therefore be of more importance than climatic cond¡t¡ons alone. However, harsh climatic

26 Chapter t conditions fit the picture of an adverse environment, which contributes with other factors' to influence the prevalence and severity of trachomatous disease.

Socioeco no mic DeveloPme nt poverty and trachoma are often seen together. Historically trachoma was found in people who were living in crowded or adverse conditions (Dunn 1995). lt has persisted in developed countries, usually in socioeconomically deprived enclaves within trachoma-free commun¡l¡es e.g. North American lndians and Korean Tanners in Japan (Editorial Med. J. Aust. 1972).

Environmental and personal factors, which play an important role in the prevalence and intensity of trachoma, are strongly associated with socioeconomic status (Treharne 1985). A review of the N.T.E.H.P. data (Tedesco 1980) showed that the environmental indices of: living conditions, housing, d¡et, water resources, sanitation, climate, and employment opportun¡t¡es, when grouped together to form lour distinct zones of socioeconomic development, correlaled with geographic differences in trachoma morbidity. Adverse or poor conditions were associated with an increased prevalence and severity of the disease. This paralleled the pattern of infant mortal¡ty and crude death rates experienced by Aboriginal

people in the Northern Territory. Tedesco concluded that socioeconomic initiatives, in

addition to changes in government policy, were required to alleviate this burden of disease.

It has been speculated that inadequate nutrition may atfect the development of immunity

against trachoma (Watson 1969) and in a survey of Aboriginal health in New South Wales,

trachoma was lound with other diseases that were themselves associated with malnutdtion

and poor medical care (Kamien 1976 a, Kamien 1976 b) ). However, there have been no

substantial studies linking diet with trachoma.

ln Tunisia, increased economic development in one village compared to another was

associated with a decline in active infectious trachoma (Dawson et al. 1976). lt has also been

found that higher socioeconomic status within a community can be protective against

trachoma (Ballard 1983, Assaad et al. 1971). Higher prevalerrce rates of tracfþma and disease

27 Chapter 1 of greater severity have been demonstrated in households with lower than average income levels and education (Wnkler 1963, Mathur and Sharma 1970, Werner and Sareen 19771-

Factors related lo poverty and poor socioeconomic status, such as crowding (Assaad et al.

1969), family size (Jones 1975), and numberof children underten (Ballard et a|1983), have all been associated with the prevalence of trachoma. The activity of trachoma has been shown to be directly proportionalto the size of the community (Tabbara and Taylor 1988). However, improvement in any singte factor is not sufficient to induce marked reductions ¡n ths prevalence of trachoma. The overall socioeconomic development within a community is important. Often the compounding eflects of behavioural and environmental changes result in a marked reduction in trachoma prevalence (Prost and Negral 1989).

Water

The availability of water has been associated with the prevalence of trachoma since early

observations of the disease in the 19th century (Boldt 1904). An early evaluation of the role

of water, demonstrated that increasing the availability of piped water, decreased the

prevalence of trachoma because of increased utilisation for personal hygiene (Marshall 1968).

Although many studies evaluatíng the association between trachoma and waler have been

reported, few have applied vigorous scientific methodology to the invesligation (West et al.

lggg). Most studies are cross-sectional compadsons. They may be subject to rnore bias and

conlounding that the alternative intervention studies (Henneken and Buñng 1987). Either

form ol study may be subject to methodological problems including, inadequate controlfor

confounding factors, problems with grading trachoma, lack of cross checks for observer

variation, differences ¡n the age structure and size of comparative populations, and

questionable assumptions concerning water utilization. Problems have also arisen with crcss-

seclional surveys, when the survey is carried out at d¡fferent times and ditferent places (Prost

and Negral 1989).

28 Chapter 1

Several associations have been reported between water and trachoma. An increasing prevalence of active trachoma was found with increasing distance between houses and their water suppty (Telsch et al. 1988). The type of water supply, and possibly the water quality, was correlated with trachoma prevalence (Marshall 1968), especially in the under 20 age- group (Assaad et al. 1969). Daily bathing, when compared with infrequent bathing, was identified as protec.tive for lrachoma (Majouk 1966). ln Australia both follicular and cicatric¡al trachoma decreased with increased water availability; however the supply ol water was not

isolated from other variables (R.A.C.O. Report 1980).

A further analysis ol these data showed that trachoma was strongly correlated w¡th lhe type

and quality of water supply and sanitation. However, other variables such as climats,

dwellings, diet and emptoyment prospects were included in the analysis to create an overall

measure of living conditions in four distinct zones (Tedesco 1980).

Amongst other correlates, per capita daily consumption of water showed an inverse

relationship with the prevalence of trachoma (Kupka et al. 1968). An increased prevalence ol

trachoma was found in communities without a continuous year round supply of water,

compared with communities with this facility (Chumbley and Thompson 1988).

ln a recent review Prost analysed published data on water and trachoma (Prost and Negral

1gB9). The analysis was limited by the selection of an appropriate measure of improved

trachoma stalus e.g. a change in severity vs absolute reduction in prevalence. Thers were

also limitations because of the difficulty in controlling for confounding factors. However, by

considering only those studies that were methodologically correct, Prost was able to conclude

that improving the water supply could change the epidemiological pattern of trachoma lrom a

severe blinding endemic disease to a relatively mild, non-blinding infection.

Whilst the provision of water appears to have a central role in decreasing trachoma prevalence,

more recent work has focussed on the utilization of water. lt has been shown that when water

is scarce, it is used preferentially lor drinking and cooking rather than lor personal hygiene

(Ballard et al. 1983). The value waler has to a family is sometimes determined by the distance

29 Chapter 1 travelled to collect it, and the perception of the way in which water should be utilized in the

home (White et al. 19721. Culturally specific fac'tors, w¡th respect to hygiene and water

utilization, are important (Taylor and Sommer 1985). Further research is needed to determine water usage in trachomatous and non-trachomatous families, and the priorities involved in the

allocation ol water (West et al. 1989).

30 Chapter 1

2.1.8. Transmlsslon ln 1904 Boldt wrote "lt is enough to bear in mind that the family ¡s the main channel for the spread of trachoma, and to a certain extent the incubator of the virus'(Boldt 1904 p 126)'

Cunent authors agree that trachoma is a family disease (Darougar and Jones 1983).

The mechanical means of transmission can occur by many routes, and probably varies from community to community, reflecting different social, environmental and economic conditions.

Several mechanisms exist in endemic areas; spread by contaminated lingers, spread by face' cloths and clothing, spread from infective discharges on bed linen, spread by direct contac't between parents and children and between children, fly transmission, and possible spread by respiratory droplets, gastrointestinal and genital contact (Treharne 1985). Dogs have been investigated as an animal reservoir lor Chlamydia trachomatis, but a link to human infection could not be found (Rhodes et a|.1987).

It has generally been accepted, although not specifically proven, that these are the principal routes of transmission. ln fact, speculation about the role of flies occurred as early as the

1930s (Macoallan 1931). Later, contributors to the trachoma literature were convinced that eye-seeking flies, which otten cluster about the eyes, nose and mouth of any person with an ocular or nasal discharge, were an important vector in the transmission ol Chlamydia

trachomatis (Jones 1975). ln 1981, a laboratory experimenl proved that fly transmission could occur (Forsey and Darouger 1981). ln Tanzania a study of hygiene factors and trachoma found that household fly density was a risk factor for inflammatory trachoma and the presence

of fties on the face of a child with nasal discharge was shown lo increase the odds of

developing trachoma almost two fold (Taylor et al. 1987). Taylor also found that acltive

trachoma in children was significantly more common in households with high fly numbers (>5

per house), compared to a range of fly scores less than this number (Taylor 1988). ln a recent

study of the impact of cattle on the prevalence and severity of trachoma found that cattle

droppings, by increasing the density of llies, contributed to the transmission and prevalence

of trachoma (Sole 1990). However, with respec't to the Taylor study, it is possible that

31 Chapter 1 trachoma caused the nasal discharge and attracted the flies (West et al. 1991). Further controlled lield trials have not been reported, and the aclual importance of this route remains speculative.

Extraocular sites of Chlamydia trachomatis play an important role in transmission (Schachter and Dawson 1981). Auto-inoculation, ¡f proven, has practical implications ¡n trachoma trealment and control, especially in the younger age'groups.

2.1.9. Trachoma Treatment and Preventlon

The early Egyptians, Greeks and Romans used copper to treat trachoma. This was continued through the ages, and before the introduction of sulphonamides in 1938, patients had their everted eyelids scoured daily with a pointed crystal ol copper sulphate held in a wooden holder. This was often continued for years until the disease became quiescent. lt was a painful and sometimes fruitless treatment; however there were more drastic measures.

lnoculation of the eye with Gonococcus was practised in the early 19th century in cases with

severe pannus. Later treatment methods used s¡lver n¡trate, lotions of mercuric chlodde, or

subconjunctival injections of cyanide of mercury. X-rays were also used, all to ach¡eve

resolut¡on of the chronic inflammation (Duke-Elder 1965). Surgical correction by excision of

the diseased areas was prac;tised for many years and vadous operations were advocated

(MacOallan 1913).

The introduction of antibiotics, and a better understanding of the disease, heralded a new era

in the treatment of trachoma. tn Australia, Flynn used sulphonamide drugs: Sulphatrial, 6

tablets or I grains to begin with, then 3 tablets twice a day for two weeks and Aureomycin

ointment (1%)topically, twice orthree t¡mes a day. He repeated the course twice at intervals of

2 weeks, and suggested the use of mass campaigns based on education and prevention as a

means of controlling or eradicating trachoma (Flynn 1957).

32 Chapter 1

Treatment and prevent¡on today can be divided into five levels of intervention, and uses the basic elements of Flynn's approach:

1. Primary Prevention

Trachoma is a disease which is preventable using a primary health care approach.

Hislorical and anecdotal data suggest that environmenlal factors and hygiene

practices play a major role in Chlamydia lransmission and endemicity (Taylor and

Sommer 1985).

There has been little sociomedical research in the area of primary prevention (Dunn

1gS5). Observations that trachoma disappeared in Europe and rnost of Australia pdor

to intervention programs and that trachoma is unequally distributed in hypoendemic

areas, clustering in family or household groups, indicate that the use of a public health

approach would be useful (Taylor et al. 1987). Strategies that were non-medicalwere

lound to be effective in controlling trachoma in South Africa (Kok 1983).

ln the prevention of trachoma three areas of culture have been identified as impodant

(Hollows 1gB5). The first involves the 'major health haróruare"; the provision of water,

adequate living space, elevated separated and ventilated sleeping spaces, above

ground surfaces that can be cleaned and are free ol animals and potential insect

vectors. The second describes "softwars' which includes cultural change and

behaviour modification and programs that are understood and designed by the

participants. The third involves the 'client group" which implies an understanding of

the people, a comm¡tment lo the group's welfare and advancement, and an

appreciation of other health-related effects that may precede or follow the

development of a trachoma control program. This would involve the support and

invotvement of the client group at all times. Another approach emphasises the

pdmary prevention of trachoma which inctudes health education, promotion of self

teaching, and a recognit¡on that socioeconomic development should be

accompanied by behavioural and lifestyle changes. Three factors are crucial for

33 Chapter 1 success: compliance with long term behavioural modification, integration with health care services, and eye health education (Thylefors 1985).

Both the approach outlined by Hollows and that by Thylefors have overlapping features; however, the latter emphasises the individual whereas the former incorporates physical, q.¡lturaland community aspects in an overallscheme. ln Third world countries any system that ¡s established should be affordable, acceptable and accessible to the people (Sutter and Ballard 1988). Furthermore a clearty stated health pnticy needs to be formulated. This would include detailed activities and general guide-lines (Mburu and Steinkuller 1983). Currently, for example, demands for ophthalmic treatment serv'rces in Africa far exceed resources

(Mburu et al. 1983). To fulfilthis type of demand apprcpriate technology and training of health care workers is needed (Sandford-Smith 1984). During the 1980s there had been a steady increase in the incorporation of non-physician trained workers ¡nto eye care programs (Pizzarello 1990). Training programs for health workers have been described for countries such as Kenya (Sheffield 1983), and Australia (Brian et al.

1990). A distinction between leaching routine clinical procedures and teaching

strategies for problem solving has been made (Gooóruin 1983). The success of using

health workers in trachoma control programs has been demonstrated in at least one

study, a significant reduction in the severe grades of trachoma occurred with the use

of primary heatth care workers and community action (Anderson and Bentley 1986).

ln Australia, workshops and courses have been establ¡shed to tra¡n rural health

workers in ophthalmic assessment, primary treatment and sometimes definitive care

(Brian et al. 1990). Lastly, the pronntion of eye health by advertising, leaflets, posters

and community meetings, has been demonstrated to lead to successful cooperation

by individuals with trachoma controlprograms (Cedrone et al. 1987). '

34 Chapter 1

2. Secondary Treatment

Secondary treatment is the use of medications to eradicate disease once it is established. ln 1981, the World Health Organisation recommended twice daily treatment with tetracycline ointment for 5 consecutive days, once a month for six months (Dawson et al. 1981 a ). A controlled trial demonstrated that a significant reduction in trachoma was produced. However, the etfect was short-lived (Dawson et al. 1981 b ).

A recent randomized controlled trial of trachoma treatment in the Northern Territory compared antibiotic eye drops, eye washing and no intervention. There was a 33% reduction in infec{ive trachoma in those treated with eye drops, compared wrlh a 2T/"

and 24o/o reduction in the eye washing and control groups respectively. This significant result was based on screening2852 children in 4 commun¡t¡es, 1164 of whom were found to have trachoma and were enrolled into the study (Peach et al.

1gB7). However, the overall program was criticised because under tr¡al conditions compliance was only 60%, the etfects were shorl term, and therefore only palliative,

and Chtamydia trachomatis from other sites was not eliminated. Also the pool of

infection which existed in pre-school children, who were heavily infected, was not

reduced and there was a possibility that res¡stant strains could have emerged

(Matthews 1987).

Recent evaluation of the efficacy of oral antib¡ot¡cs in the treatment ol Chlamydia

trachomatis infection in Aboriginalchildren is lacking. Comparisons between topical

and system¡c treatment have, however, been undertaken elsewhere. A stratified

double btind controlled field trial in a village in Southern lran demonstrated no

dífference in the prevalence of trachoma atler four weeks of treatment. The treatment

consisting of the administration of topicaltetracycline twice daily for 7 days once per

month, versus oral doxycycline Smg/kg once monthly (Darougar et al. 1980). After

one year there was a marked reduction in the prevalence and intensity of trachoma for

35 Chapter 1 both regimens. Oraldoxycycline; however, was considered more practical and less expensive.

On the West Bank in lsrael, a double blind stratified comparative trialwas undertaken of intermittent, family based therapy lor trachoma, using oral doxycycline, oral sulphametopyrazine, or topical tetracycline eye ointment. lt la¡led to dernonstrate any statistically significant ditference between the 3 groups treated in the two villages

(Chumbley et al. 1988). However doxycycline, given once weekly for three weeks was the easiest to adm¡nister, although compliance was high for alltypes of treatment.

A further study has been undertaken to examine the effect of treatment regimens.

This study suggested that systemic antibiotics should only be given to children with severe or moderate intensity disease, and those individuals at s¡gn¡f¡cant risk of

blindness (Dawson and Schachter 1985).

There is much debate about the most appropriate form of systemic therapy. ln 1982,

Dawson compared a topical eye antibiotic and oral erythromycin. Although

erythromycin proved to be marginally more etfective in eradicating trachoma, aclive

disease did persist in many children (Dawson et al. 1982). A subsequent study of

medication compliance revealed that lack of compliance was probably the cause for

this difference. Erythromycin had been administered four times daily, but to achieve

compliance this was reduced to tw¡ce daily during school hours. Serum levels ol

erythromycin were found to be negligible in the early morning, indicating rapid

clearance of the drug as being the reason for treatment failures (Dawson et al. 1982).

An alternative to erythromycin ¡s tetracycline or its longer act¡ng derivative

minocycline, which requires only one daity dosage. A double blind field trial of oral

minocycline vs tetracycline ointment demonstrated equal etficacy atter three weeks.

However, minocycline was superior at evaluation after one year (Tabbara 1990). The

once a day dosage required for minocycline was thought lo ensure compliance.

However, the tetracyclines are contraindicated in children under I years of age

36 Chapter 1 because of the possible adverse effecls including: (i) deposition of tetracycline in calcifying tissue such as teeth and bone, causing permanent discolouration in the former and stowing bone growth in the latter and (ii) photosensitive dermatitis (MIMS

Annual 1990). This is the group which is targeted in most control programmes.

The comparative costs and benefits of using tetracyclines in trachoma endemic areas were analysed. lt was concluded that oral doxycycline was at least as good as topical treatment, and that the ease and feasibility of once a day treatment was attractive fôr public heatth programs. By limiting the adminislration of systemic doxycycline to those children with severe intensity disease, adverse side effects would be limited to those most at risk of blindness, such effects would occur in only a small percentage of children, also it was cost effective to combine intermittent systemic treatment with an ongoing mass campaign of topicaltreatment (Dawson and Schachter 1985).

Sulpha agents have been used ln the past (Bietti 1957) and were used extensively in the N.T.E.H.P. survey 1976-1979 in Australia (R.A.C.O. Report 1980). They are now

not generally recommended because of adverse reactions including allergies, renal

side effects and the severe Stevens'Johnson syndrome.

The q¡rrent utilization of antibiotics can be summarised as follows:

(1) lntermittent topical therapy, which can significantly reduce the

intensity of eye infections due to Chlamydia trachomatis (Reinhards et al.

1968), has been used in numerous ant¡-trachoma campaigns (Thylefors

1985). lt is viewed as a prophylactic treatment because although the blind¡ng

complications of the disease are prevented, trachoma may still exist in a mild

form in a community (Kupka et al. 1968). lt has the secondary effect of

reducing the incidence of bacterial super-infection which may lead to a

reduction ¡n the bl¡nd¡ng complications of trachoma.

37 Chapter 1

(21 The use of systemic antibiotics in individuals who either do not

respond to topical treatment, or who have blinding complications, e.g.

trichiasis or pannus. Tetracycline or doxycycline is used in adults, but not in

pregnant women because of possible teratogenic etlects (MIMS 1990), when

erythromycin should be used (Schachter et al. 1986 b ), and erythromycin is

generally used in children.

3. Tertiary Trcatment

Tertiary treatment involves systematic case finding and corrective lid surgery' ln

endemic regions mild entropian with trichiasis may be very common, the result of

many years of infection. The corrective procedure is simple and can be performed by

auxiliary medical personnel. A controlled tr¡al of five operalions for the correction of

trichiasis has been undertaken by Reacher et al. (1990). They found that bilamellar

tarsal rotation and tarsal advance were etfective operalions for maior trichiasis, but that

tarsal advance was inetfective where trichiasis and lid closure defects coexisted.

Moderate cicatricial entropian with or without trichiasis, secondary blepharospasm and

secondary brow ptosis require more extensive assessment and corrective

procedures which have recently been evaluated in Saudi Arabia (Nasr 1991).

The progression of trachoma is such that even after active infec{ive disease has been

controlled, a lag period of 10-20 years may exist before those whose lids were scarred

in chitdhood reach adulthood and begin to experience difficulties with trichiasis

(Dawson and Schachter 1985). Therefore tertiary strategies must be ma¡ntained for a

long period.

4. Vaccination

Chtamydia trachomatiscauses not only oq¡tar lrachoma but also genital trac't infection.

It is one of the most comfrþn sexually transmitted pathogens in western developed

38 Chapter t countries, and can ult¡mately cause infertility in women from chronic salpingitis.

Controt ol Chlamydia trachomatis is thus of importance to public heatth globally.

A major search lor a trachoma vaccine has been in effect for thirty years. Although some short term benefits were demonstrated with live unit vaccines (Collier 19741' some vaccines produced hypersensitivity, inducing more severe disease

(Monnickerdam and Pearce 1983). lt is thought that the difficulty in developing an eflective vaccine relates in part to the poor immune response to Chlamydia infection.

Ocular infection wtltt Chtamydia trachomatis induces both humoral and cell mediated immunity (young and Taylor 1984). Although some partialprotection is conferred with these responses, it also may be responsible for the tissue damage and pathology seen in trachoma (Taylor and Sommer 1985). An autoimmune mechanism may be responsible (Abu El-Asrar et al. 1989). lt is clear that an effective vaccine should separate the sensitising from tho prote¡n anligens, it should be a subunit of the

Chlamydia organisms, and that the problem of serotype-specific immune responses needs to be addressed (Schachter f985)'

A'killed' vaccine would be unlikely to confer extended immunity. A desirable goal would be 4-5 years of protection; this period matches the t¡me of peak incidence in the pre-school age-group. Protection during that period could prevent the

development of changes which lead to blindness in later life. Any vaccine would be

potentially elfeclive against genitat infection, conferring protection at least to the

effects of salpingitis, if not from infection of lhe mucous membranes ol the lower

genital tract (Schachter I 985).

5. Mass Campaþns

since the late 1950s a number of trachoma control programs have been undertaken

(Thylefors 1985). Most mass campaigns have relied upon topical tetracycline,

although some have utilized oral antibiotics of various types (R.A.C.O. Report 1980,

Darougar et al. 1980, Dawson et al. 1981 b ). Common to mosl campaigns have been

39 Chapter 1 five main strategies : (1) early detection of cases, (2) prophylactic treatment on either a mass or selected basis, (3) topical and systemic treatment for long periods for severe infection, (4) surgical treatment of trichiasis, and (5) health education (Thylefors

1 985).

It has been conjectured that the costs and benefits ol topical tetracycline treatmenl

are low, whereas benefits and costs of system¡c antibiotics are higher (Taylor 1987 a ).

Studies evaluating the success of campaigns are mixed. ln Southern Morocco the

overall prevalence of trachoma was unaffected in a mass campaign incorporat¡ng the

above four steps (Kupka et al. 1968). However, a larger proportion of cases were at

the ',healed slage lV level", and in one region there had been a decrease in the active

I to lll stages in the younger age-groups. ln Tunisia, an evaluation of an ongoing

campaign demonslrated a decline in the prevatence and severity of the d¡sease

(Dawson and Schachter 1985).

ln Austral¡a a coordinated mass campaign used oral co-lr¡moxazole in a method

outlined by Jones (1978), of an attack phase followed by a consolidation and

surveillance phase. A substantial reduction ¡n the prevalence of trachoma follicles in

children was demonstrated. Factors such as observer variation, natural remission with

increasing age, seasonal changes and improvements due to changes in living

conditions were taken into account (R.A.C.O. Report 1980)' However, mass

campaigns utilising systemic ant¡b¡ot¡cs are expensive and difficult to complete. ln

addition, as with the use ol sulphonamides, adverse reactions to antibiotics increase

with indiscriminate use. By contrasl topicalmed¡cat¡on is effective in eradicating oqJlar

Chtamydia trachomatis, is cheap, easy to use and has feurer side etfects.

More recently campaigns have focussed on trachoma as a community or family based

disease (Hollows 1985). A public health approach addresses the underlying

socioeconomic and cultural factors which foster the pers¡stence of trachoma in

communities. The 'natural" etimination of trachoma lrom Europe (Jones 1975) and

40 Chapter 1 white Australia (Mann 1957 b ) preceded the ¡ntroduction of modern antibiotics, underscoring the need for socio-cultural and economic changes in communities if treatment campaigns are to be ultimately successful.

The discovery of a suitable vaccine to augment the immune system without damaging ocular structures woutd be desirable. However, ¡t ¡s acknowledged that the eradication of trachoma at present rests in the area of prevention and intervention

(Tabbara 1990). Mass campaigns and public education strategies are urgently required in many developing countries.

41 Chapter t

Summary

Trachoma is an important cause of visual disability throughout the world. The causative organism is Chtamydia trachomatis, an intracellular bacterium which not only infects lhe eye, but also may infect other mucus membranes and organs in the body. Trachoma may be acquired at any stage in life, but in endemic regions infection often begins in the lirst one or two years. Recurrent infection may produce the tarsalplate scarring and lid c¡ntracture which

teads to trichiasis and cornealopacity in later life. lt is thought that trachoma is a mullicyclic

disease w1h relapses and remissions over many years; it is also, as already mentioned, a rn¡lti-

system disease, which may explain treatment failures when only topical treatment is used.

Trachoma thrives in crowded and unhygienic conditions and can disappear with

socioeconomic development. Although all ages can be infected, ¡nflammalory disease

predominates in the young and cicatricial disease in the old. Women appear to have higher

rates of inlection than men. Cutturalfaclors parl¡cularly water usage and hygiene have been

identi¡ed as important ¡n ¡ts epidemiology. Public health measures are the key to the

prevention of trachoma, secondary treatment w¡th either intermittent topical antibiotics or

systemic antibiotics ¡s the second line of defense, with tertiary procedures such as eyelid

surgery, being reserved for those in whom the trachoma has progressed to its blinding

stages. Currently no effective vaccination exists and so mass campaigns, which have had

mixed success in the past, willonly be used on a limited scale where funds and comm¡tment

are available.

42 Chapter 1

1.3.2. Cataracts

Cataracts are an important cause of visual loss in all racial groups. Although risk factors have been identified, effective preventive strategies have not been developed. For the Aborig¡nal population, an understanding of cataract epidemiology is important, to ¡dentify those at ñsk and to be aware of the factors that lavour people to seek and receive treatment when it is available. This understanding is also important for the design of strategies to reduce the prevalence of blindness due to cataracrt.

The lens of the eye focuses incoming light on to the retina. When the lens loses its normal transparency a catarac't is said to be present. The location, distribution and density of any lens opacity determines whether the cataract impairs vision. Most are bilateral, although the rate ol

progression in each eye is rarely equal. The most comrþn cause of cataract ¡s ageing, other

causes include: genetic predisposition, congen¡tal factors, metabol¡c imbalances, trauma to

the eye, and eye toxins (Vaughan and Ashby 1983). Approximately 17 million people

worldwide are binocularly blind because of cataracts (Dawson and Schwab 1981), and it is the

leading cause of blindness in the elderly (Hyman 1987).

The formation of a cataract occurs with changes in the physiology and structure of the lens.

Lens edema, protein alteration, necrosis and disruption of the normal continuity of the lens

filaments are characteristic. During the development of a cataracl several stages are passed

through each characterized by different physiological and structural changes. An immature

cataract is only lightly opaque, a moderately mature cataract is edematôus, a hypermature

cataract is relatively dehydrated, the lens opaque and sunounding capsule wrinxled (Vaughan

and Ashby 1983).

Visual acuity testing is the best ctinical measure of cataract opacity, the decrease in visual

acuity is generally in proportion to the density of the cataract. There are, however, individuals

in whom the reduction in visual acuity is out of proportion lo the degree of opacilication, this is

thought to be due to distortion of the visual image by a partially opaque lens. ln contrast,

43 Chapter 1 some people have marked opacity of the lens but are still able to carry out their normal activities.

The Epidemiology ol Cataract

Several difficulties arise when considering the epidemiology of cataract. Problems with

delinition occur because there has not been standardization in morphology (shape, size and

density) of the lens change in cataract between studies. Assessment of these changes is

often subjective. Secondly, changes in visual acuity which reflect cataract development may

be assessed differently between surveys, resulting in survey ditferences due to methodology

(Leske and Sperduto 1983). Thirdly, methods ol examination e.g. direct ophthalmoscope vs

slit lamp, and the selection of sample groups atfect the prevalence values as well (Sommer

19771. Lastly, groups that are selected for sampling have also varied; either from selected

communities, or age-groups, or blindness registers, or randomized surveys of entire

populations (Leske and Sperduto 1983).

ln determining the descriptive epidemiology of cataract several sources may be used

including:

1. Blindness reg¡sters from different countries have been used; Canada

(MacDonald 1965), Great Britain (Sorsby 1966) and the USA (Kahn et al. 1977).

Factors such as differential registration of blind individuats and the ditficulty of

comparing blindness rates between countries affect interpretation (Leske and

Sperduto 1983).

2. Prevalence data is available from population based studies;for example, the

Framingham Eye Health Survey. The data, however, was limited to individuals who

lived in a specific community, and who had previously part¡c¡pated in a heart study

examination (Leske and Sperduto 1983).

g. Data drawn lrom attendance at outpatients (Batterbury et al. 1991).

44 Chapter 1

Risk Factors lor Senile Cataracts

Several risk factors for age related or senile cataract have been identified in survey and case control studies. The factors can vary for each category of age related cataract viz nuclear corticat, posterior,subcapsular and mixed cataracts (Leske et al. 1991).

1. Demographic Factors lncreasing age has been demonstrated in numerous studies to be related to an increasing prevalence of cataract (Leske and Sperduto 1983).

A consistent difference between males and females has not been found for age-specific prevatence rates of cataract. Although studies have shown a higher prevalence of cataract in women, for instance in the Framingham National Health and Nutrition Evaluation Survey 1971'

1972 (Hiiler et al. 1983), and a study in Edinburgh of four hundred and eighty seven people over the age of sixty seven (Milne and Williamson 1972), a case control study of blindness in a rural Aboriginal population found that cataracts occurred as commonly in males as females

(Taylor 1980 a ).

Variations in the prevalence of cataract throughout the world have been demonstrated.

Comparison studies have found that lhe age-adjusted prevalence of cataract for an area of the

Punjab ptains in lndia (Chatterjee et al. 1982), was three times higherthan that obtained in the

Framingham study. Differences may be dus to genetic or environmental factors; however,

variation in survey melhodology or the availability of medical services, particularly for

comparisons between developing and western nations, may account for the differences

found (Leske and Sperduto 1983).

Two recent case control studies, The Lens Opacity Case Control Study (Leske et al. 1991)

and The ltalian-American Cataract Study Group (1991) have identified; or given further weight

to several factors that affect the occurrence of age related cataracts. Table 1.3'2.1

summarizes the finding of these studies.

45 Chapter 1

Table 1 .3.2.1: Risk FaCtors for Cataracts

Type of The Lens Opaclty Itallan-Amerlcan Cataract Cataract Case Control Study Group

Decreased lncreased Decreased lncreased Risk Risk Risk Risk

posterior increased low education arthritis low education subcapsular vitamins diabetes family history stero¡d use

nuclear increased low education arthdtis low education vitamins non-professional current smoking

cortical increased low education adhritis low education vitamins lemale sex female sex non-white family history sunlight

mixed increased low education arthritis low education vitamins diabetes family history sunlight

2. Vitamins

The Lens Opacity study demonstrated that individuals who had high values for an index ol

antioxidant vitamin intake were protec'ted against cataract formation but the ltalian-Amedcan

Cataract Group, with identical methodology, failed to find an assoc¡ation. Both findings were

valid only to the extent that current nutrient intake reflected past eating pract¡ces.

Further confirmation of the association between cataract and nutritional intake has been found

in a survey conducted in lndia. The study demonstraled that persons with a better nutñtional

intake had a lesser risk of cataract. The measures of dietary intake were; however, closely

correlated so an individual protective vitamin could not be identified (Mohan et al. 1989).

Some studies have; however, shown no correlation with specific vitamins such as thiamine,

pyridoxine or vitamin E (Bhat 1987).

46 Chapter 1

3. Sunlight

Sunlight, or Íþre particularly near ultra-violet (UV) light has been demonstrated in-vivo and in- vitro studies to induce opacity of the crystalline lens (Zigman 1973, Pirie 19721.

Epidemiological studies in the 1970s, indicaled a relationship between cataract prevalence and high sunshine intensity (Zigman et al. 1979, Hiller et a\.1977). ln Australia, a case control study of 350 Aboriginals aged 30 and over demonstrated a strong association betwee.n cataracts and latitude, increased exposure to sunlight and of high ultra-violet-B (UV-B) radiation (Taylor 1980 b ). A more extensive survey of 64,307 AborÍginals revealed a pos¡t¡vo corretation between the prevalence of senile cataract and levels of climatic UV radiation

(Holows and Moran 1981). An equally large survey in Nepal (30,565 lilelong residents), demonstrated that people who came from sites with an average ol twelve hours of sunlight exposure per day, had 3.8 times the risk of catarac{, compared to those from sites with an average of only seven hours of exposure (Brilliant et al' 1983 a )

More recent studies using sunlight exposure, as ir¡ferred from interview data on residency and time spent in the sun, have shown a slight relationship between increased exposure and

irrcreased risk of cataract (Collman et al. 1988).

A study using retrospect¡ve data from the 1971-1972 National Health and Nutrition Evaluation

Survey and the average daily UV-B radiation at the site of the examinat¡on suggested a

positive correlation with corticalcataracts (Hiller et al. 1986).

Both the Collman and Hiller studies suffer from the methodological problem of determining

the true life-time exposure of an individualto UV-B radiation, as opposed to the reported or

assumed exposure.

Further studies have been undertaken and these have attempted to quantify an individual's

lifetime exposure by taking a careful history. ln a study of Maryland watermen, a positive

correlation between sunlight and cortical cataracts was demonstrated (Taybr et al. 1988). A

case-control study of individuals operated on for poster¡or subcapsular (PSC) cataracts in

47 Chapter 1

Maryland (Bochow et al. 1989), suggested that UV-B exposure may be an important risk factor for psc cataracts. By contrast, in the Lens Opacity case control study, no correlation was found for cortical or mixed calaract, and only a slight association for nuclear cataract. The authors thought that the urban population studied would have had limited sun exposure, and they were cautious about using self-reported data on sun exposure (Leske et al. 1991).

A similar study in Parma ltaly, did demonstrate an association between corticalcataracts and sunlight exposure. Sunlight exposure was est¡mated using three variables: work location in the sunlight, teisure time in the sunlight, and use of a hat in summer. Surpdsingly, use of a hat was not protect¡ve and may in fact have been a marker for increased exposure (The ltalian-

AmerÍcan Study Group 1991). Although there is evidence suggest¡ve of an association between UV exposure and cataracts, Harding (1982) and Harding and von Heyningen (1987) have speculated that because conditions associated with sunlight exposure (e.9. pterygium

and c¡matic droplet keratopathy) are not associated with cataract, and because in-vitro studies

are not supportive of the association, that other factors are rnore important.

4. D¡abetes

There is biochemical evidence that elevated sugar levels which occur in diabetes, leads to an

accumulation of sorbital which can damage the lens (Kinoshita et al' 1979). An

epidemíological review of cataracls (Leske and Sperduto 1983), summarised that although

some studies based on extracted lens showed an increased dsk among diabetics (Hiller and

Kahn 1976), others based on population prevalence surveys did not (Waite and Beetham

1 e35).

It was hypothesized that because diabetics were more likely to seek medical atlention, they

woutd probably have their cataracts identified more otten than non-diabelics, resulting in an

ascertainment bias (Sommer 19771.

More extensive surveys have demonstrated a positive correlation between diabetes and

cataracts (Ederer et al. 1981). Studies havs atso identified that the postedor and cortical

48 Chapter 1 cataract types were more common in diabetics than nuclear types (Hiller et al. 1986).

Confirmation of this relationship occurred in the Lens Opacity Case Control Study (Leske et al.

1991); where diabetes was associated with cortical and mixed cataracts, but not nuclear cataracts. The authors suggested that lhe lack of a relationship was due to the pathogenic

mechanism, which they speculated predominantly affec{ed the lenticular zones of the lens.

Several other factors have been identified as playing a role in the development of sen¡le

cataract. High doses of ionising and infra-red radiation are known to produce cataracls (Miller et al. 1967, Lerman 1980). Some drugs have cataraclogenic potential including

corticosteriods, phenolhiazines, enzyme inhibitors, metals, organic compounds, and cancer

chemotherapy agents (MIMS 1991). Blood pressure (Kahn et al. 1977), family history (The

Itatian American Study Group 1991), myopia (Belkin ot al. 1982), eye colour (Collman et al.

tgBB), education (The ltalian American Study Group 1991), and metabolicfactors (Clayton

i9B4) have all been identified as potential risk factors for catarac"ts. Some of lhese, such as

the role of dehydrational crisis or use of aspirin, are contentious.

A history of dehydrationalcrisis has been demonstrated in some stud¡es to be associated with

cataract formation (Mianassian et al. 1989), and laboratory studies have given biological

plausibility to this hypothesis (Harding and Rixon 1981). However, other stud¡es in similar

populations have failed to support this hypothesis (Bhatnagar et al. 199f ).

The use of low dose aspirin has been shown to be of no benefit in some studies (West et al.

1987), but observational studies support its role (Mohan et al. 1989). Methodological

problems have been identified in both the West and Mohan sludies, and the findings of a

Physicians Health Study, although null lor cataract development and not statistically significant

for cataract extraction, were compatible with a small protective etfect of aspirin therapy

(Seddon et a|.1991). The authors concluded that further investigation was warranted with the

Health Study's cohort. They also argued, that a possible reduction in the number and

therefore cost of cataract extraclions in the U.S. using drugs (by slowing the development of

49 Chapter I cataract), should be balanced against the potential side effects of any agent used. This would require careful evaluation of dsks and benefits before a public policy was formulated.

Blindness due to Catarad

Cataract has been identified as the leading cause ol blindness ¡n tho world (Sommer 1988).

Devetoping nations bear the brunt of this burden (Schwab 1990 a ), and in Afdca the shodage of manpower and material resources is recognized as the cause of an increasíng backlog of treatable blindness (Foster 1987), a backlog that grows by 160,000 per year (Steinhuller

1s83).

Various strategies have been proposed to help ovsrcome this problem. These include: fostedng the politicalwi[ to support a blindness prevent¡on progfam, obtaining resources, and using cost efficient techniques,which might include using non-physic¡an ophthalmic practitioners to perlorm cataract operations (Sctwab 1990 a ).

Several factors are important in consideñng any program (Foster 1987). Firstly the pat¡ent

must be motivated to seek help. Past experience or the experience of friends or relatives with

cataract surgery may affect this, as woutd lengflhy traveltime, expense, and unfamiliarity with

physicians. Secondly three steps should occur in the provision of services within a program,

screening, the selection of those likely to benefit from surgery, and surgery that utilizes

available resources. Thirdly issues such as the availability of manpower, the use of materials,

the mobility of patients, and lhe management of any progEm are all¡mportant.

Although the shortage of ophthatmological manpower is acute in developing countries (Smith

1990), backlogs of patients can occur in places such as the UK due to an ageing population

(Batterbury et al. 1991). Local ¡n¡t¡atives, expansion of daycase surgery and managedal

changes are seen to be important in ctearing this backlog (Drummond and Yates 1991).

Another issue that can arise even when surgery is available, is underutilizat¡on

50 Chapter 1

(Venkataswarmy and Brilliant 1981). Geographical isolation, education and economic cost have been identified as contributing to this (Brilliant and Brilliant 1985)

51 Chapter 1

Summary

cataracts are an ¡mpoftant cause of blindness throughout the world. several associations for

cataract have been found including increasing age, female sex, diabetes, lower education,

and increased exposure to U-V radiation. A decreased risk of cataraci has been noted in

individuals with increased vitamin intake and those having arthritis'

Although these associations may be viewed as risk factors, intervention studies have not

been concluded, and so prevention must rely on surgical cataract extraction. The availatility

and use of these tertiary treatments may vary, and strategies have been developed to

overcome these Problems.

52 Chapter t

1.3.3. Eye Trauma

Eye injuries are a common and often preventable. Although not a major cause of binocular

blindness in either developed or underdeveloped c,ountries, eye injuries are a maior cause of

monocular blindness in developing nations (Schwab 1990 b ).

ln Aboriginal communities the contribution of eye trauma to rnonocular blindness has already

been identified (R.A.C.O. Report 1980); however, little work has been published on using

prevention to alleviate this burden of visual disability.

Causes ot oalar iniury

The causes of ocular injury vary throughout the world, and it appears that social background

influences the aetiology and severity of ocular trauma (Adala 1983). Currently in westem

developed nations, sport, work, children at play and assault predominate as major causes

(Schein et al. 19BB). To this list could be added ocular iniuries during ¡ntrauterine life or birlh,

injuries due to travel, agricultural eye iniuries and war injuries (Duke-Elder 1965)'

Surveys in Africa have identified domestic activ¡ties as being the maior source of ocular injury

(lslar et al. 1982, Olurin 1973), the causes were: chopping wood, sharp and blunt ¡nSruments,

falls, and burns. ln a survey of eye disease in the Navajo lndians in the United States of

America, trauma from assaults was the leading cause of monocular blindness, the assaults

were often associated with alcoholingestion (Friederich 1982).

Trauma was identified as being responsible lo¡ 2.4"/" of all bilateral blindness in a Nepalese

eye survey (Brilliant et al. 1985). Bilateral blindness may also be caused by: thermal bums,

when a person falls into a fire (e.g. during an epileptic lit); road accidents, where the person is

thrown through the windscreen, and chemical burns in industrial workplaces (Roper-Hall

1 978).

The distribution of ocular injury within a populat¡on is said to be bimodal, with young and old

most affected (Tielesh et al. 1989). Although this probably relates to the timing of the iniury

53 Chapter 1 and not the crude prevalence. Men have been lound in a Stockholm study to bear thg burden of perforating injuries (Blomdahl and Norell 1984), and men attended for the treatment of eye injuries more lrequently than women (Schein et al. 1988).

Types of injury

The three most comtnon vision threatening emergencies are chemical burns, ruptured globes and hyphemas (Shingleton 1991). The first requires immediate irrigation with water followed by ophthalmic assessment; the second may be missed, as up lo 20% ol patients with ruptured globes do not have readily apparent signs of perforation;the third may be complicated by a further bleed, and 30% of patients with a hyphema have associated glaucoma, corneal damage, cataract, choroidal rupture or posterior-segment bleeding (Shingleton 1991).

Prevention

The evaluation of eye injuries, coupled with advanced surgical techniques and a better

understanding of the mechanisms of injury have improved the prognosis lor eye iniuries

(Shingleton 1gg1). However, because of the high cost of medical and surgical intervention

efforts should be direcled at prevention. This is particularly so in developing nat¡ons where

factors such as the d¡starrce to treatmerìt facilities, scarcity of eye trained statl, lack of transport

and cultural barders affecÍ treatment (Schwab 1990 b ).

ln developed countries prevention can be focussed on face and eye protecl¡on in sport

(Vinger 1981), an increasing awareness of the hazards in industry (Schein et al. 1988), and

the reduction of hazards in the home (Grin et al. 1987). ln developing countries strategiesto

reduce blindness from trauma could include: eye safety programs utilizing government

agencies,-pr¡vate volunteer organizations and health professionals (Schwab 1990 b ).

Programs to improve basic eye care facilities in rural areas have also been suggeled;as have

policy changes, such as, by-laws requiring that motorists wear seat belts (Adala 1983).

54 Chapter 1

Summary

Eye trauma is an important cause of monocular blindness throughout the world. Preventive strategies have been evaluated and ¡mplemented in developed countries. However, in thifd world nations, although the causes of eye trauma are well documented, only limited

¡ntervent¡ons have been achieved. There are also differences between developed and developing nat¡ons in the provision of advanced surgical techniques to treat eye trauma, with developing countries sometimes being unable to provide even basic accessible eye care lacilities.

55 Chapter 1

1.4 Methodologlcal Problems ln Eye Dlsease Epldemlology

Although a cross-sec.tional prevalence survey of eye health and visual impairment may seem to be a simple counting exercise, there are a number of methodologicalproblems which nu¡st be considered. By recognising their importance, procedures can be instituted to eliminate or minimize their etfect. lf they cannot be altered the results may be reported with appropriate acknowtedgement of possible inaccuracies. Only by addressing the methodological problems, which are more controllable than confounding and bias, can a study achieve its stated purpose (Hennekens and Buring 1987).

The Number of Subiects

Other factors being equal, as the number of sub¡ects increases in a study, so does the

accuracy of any estimates. This is not a l¡near relationship, and depending on the prevalence

of a condition, increasing the number of subjects beyond an accepted limit contñbutes l¡ttle to

the precision of the estimate. A law of diminishing returns means there will always be a trade-

off between the additional benefit and the etfoil required to obtain number of subjects.

CI importance in ophthalmological investigations is whether we counl eyes or subiects. Either

eyes or indivlduals may be counted and the paired eyes of individual subjects, and the

unpaíred eyes of different people must be distinguished (Ederer 1973). Put simply, the

correlation between the right and lett eye is very high for some parameters, and therefore the

results for each eye cannot be considered independently. For examplg, the trachoma stalus

of the dght eye will very likely be correlated with the trachoma status of the left eye because

both have been subject to the same physiological and environmental conditions. ln data

analysis positively correlated values provide more information when considering the

difference between paired measurements and less information when considering the sum or

average of paired measurements. Negative correlation would have an opposite etfect (Ederer

1 s73).

56 Chapter 1

The coetficient of correlation is a mathemat¡ca¡ construct indicating the degree of sameness between two parameters. lts value ranges between -1 (peíect negat¡ve correlation) through zero (absence of correlation) to +1 (perfect positive correlation). lf, for example, both lett and right eyes were pedectly positively c¡rrelated (r = +1) rþ nþre information would be gained from considering 60 pairs of eyes from 60 individuals than from

60 individual eyes, left or right, from 60 people. Alternatively, ¡f the left and dght eyes were independent (r = 0), we could consider only 30 individuals to collect the equivalent amount ôf information i.e. 60 eyes worth. Generally the correlation coefficient does not lie at these extremes, and it is evident that the degree of information provided by a pair is inversely related to (t - Q. Analysis may require comparison of left and dght eyes, or if they are highly conelated, tabulation ol only the left or right eyes may be suflicient in presenting data as the other eye adds little information. ln summary, it is important to distinguish between the number of eyes or subjects. ln ophthalmology, ocular parameters between paired eyes are usually positively correlated.

When pooling or averaging positively correlated parameters or measurements, the greaterthe

"r' correlation coefficient, the less information obtainable when compared with two independent measurements. No additional information is obtained from pooling if the parameters are perfectly correlated. However, it is sometimes desirable to tabulate left and

right eye measurements separately. Statistical methods are available to adiust for the

intraclass correlation between eyes (Rosner 1982).

Disease Definitions and Classifications

ln general clinical practice definitions of disease tend to be broad, subject to ¡nterpretation and

subject to re-evaluation over time. ln epidemiology this is unsatisfactory because national,

international and longitudinal comparisons need to be made. Standardized disease

definitions are required which rely on reproducible measures. This may mean narower

categodes and an emphasis on objective measures.

57 Chapter 1

Ophth alm¡c Obse ruation s and M e asu re me nts

Diagnosis in ophthalmology is often based on observations, which by themselves are subjective. The conditions under which an observation is made, the tra¡n¡ng and experience of the observer and the cooperation of the pat¡ent all may influence the final diagnosis

(Leibowitz et al. 1980). Reproducibility and good observer agreement have been established for a number of ophthalmic measures but there are still problem areas, and even when standards exist they are generally not lollowed (Ederer 1983)'

Visual Acuity Testing

Centralto any eye survey is an assessmer¡t of visual acu¡ty. Both examiner and subiect can have an etfect on the accuracy of this test. Double masking, when feasible, can help eliminate the b¡as (Ederer 1975).

Examiner masking when visual acuity is an outcome variable, is possible in therapeutic trials or controlled studies, but ¡s not applicable in general survey epidemiology because a pailicular oulcome can not be anticipated. Masking in a survey would be difficult, and would involve the examination of people not inctuded ¡n the survey. This would create additionalwork without improving the accuracy of the survey. Standardising pocedures and equipment (see below), then evaluating ditferences between examiners would be a more etfective strategy to pursue.

Subject masking is important only in trials when study treatments are identifiable by the

subject. Subject compliance, etforl and deception cannot be entirely eliminated by study

design. lt can be minimized by encouraging maximum effort in reading lines, and by

continu¡ng lhe examination until two or rnore mistakes are made per 8 letter line (Aiello et al.

19741. Repeat testing to reveal inconsistency, at the same or a different time, and adequate

explanation of the importance of the examination, also help to reduce subiect bias in a survey.

Physical constraints have been identified as additional sources of inaccuracy (Ferris and

Ederer 1979):

58 Chapter 1

1 Location of the light for the eye chart, or front, rear (through glass), or

projected image.

2 Amount of chart illuminatíon.

3 Distance of the patient frcm lhe chart.

4 Degree of black-white contrast between leüers and background.

5 Degree of reading dítficulty of letters (some letters are more d¡ff¡cult than

others).

6 Printing style of the letters. ln survey work it would appear that consistency of an approach to visual acuity testing is as important as limiting the physical sources of inaccuracy. surveys in remote areas, although acknowledging ditficutties, have not attempted to evaluate the magnitude of any inaccuracy. lt would appear each should be examined independently and a broader range of error

accepted, thus reflecting the constraints of the physical environment (Ederer 1975).

Re produ cibi lity of Obse ru ati o ns

lnlraobserver and inter-observer variation are ¡mpoftarìt in survey work which is not perlormed

under controlled conditions, and may involve diverse locations, several examiners and an

extended time period. ln data collection it is important to distinguish between systematic and

non-systematic error. Non-systematic error, tends to average to zero over large samples,

whilst systematic enor can lead significant bias in results (Cochrane and Gemmel19T7l.

Only recently have ophthalmic studies recognised that vadat¡ons in clinical assessments can

occur either between observers or between observat¡ons made by the same observer on

different occasions. ln trachoma stud¡es an awareness of the importance of systematic

observer variation occurred in the 1960s. The first study of the subiect occurred in Taiwan

(Assaad and Maxwell-Lyons 1967). They concluded from the results of a survey which

59 Chapter 1 examined 35,000 persons, that variations were inevitable if based on clinical examination alone. lt was better to have two observers and it was possible to assess and reduce systematic observer bias. Other studies to account for observer variation were conducted by

Hollows and Graham (Gibson and Sanderson 1980) and the Framingham Eye Survey of 1975

(Kahn et al. 1977)).

Trachoma has been the subject of frequent population-based surveys; however only a few

have reported strategies for reducing inter-observer and intra-observer error. Brilliant

reviewed four surveys and presented Table 1.4.1 (Brilliant et al. 1983 b )

60 Chapter 1

Table 1.4.1 : lnter-observer agreement on the diagnosis of trachoma:

comparison of results f rom four population surveys

o/o Survey Trachoma No. of No. of Kappa

(reference) measure subjects examiners agreement

Taiwan Absent/ 574 2 80.5 0.59

(Assaad 1969) present

Morocco Nonactive/ 200 2 77.0 0.53

1 (Kupka 1968) active

Australia Absent/ 172 2 83.1 0.53

2 (RACO 1s80) present

Nepal Absent/ 507 5 80.5 0.60

(Brilliant 1983) present

- (see 1 . Nonactive trachoma = slages O, D and lV; aciive tracfþma = S1â9êS I lll

Kupka for definition of stages).

2. Folliculartrachoma.

Although the % agreement appears to be excellent, the Kappa (K) statist¡c (see below)'

showed only fair to good concordance in trachoma grading. This may reflect the complex¡ty of

61 Chapter 1 the old trachoma grading system, lt may be concluded that because the K values were similar, with consistency between surveys conducted in different locations with different grading personnel, that study design, including standardization procedures, could not overcome the inherent problems with the complex grading system. The new five sign trachoma grading system was designed to overcome these deliciencies (Taylor 1987 c ).

The Kappa stat¡st¡c is a standardised measuro of agreement which corrects observed agreement between examiners for that expecled by chance. A three category ranking provides a measure of agreement (Fleiss 1981).

K = 1.00 - 0.75 excellent agreement

K = 0.74 - 0.40 fair to good agreement

K = 0.39 - 0.0 poor agreement

The Kappa statistic was used in the 1989/90 survey to assess the degree ol agreement for

the presence or absence of trachoma, and the effect ol magnification vs non-magnificalion on

trachoma grading (section 5.2.2 and 5.2.3)

C ro ss- secti o n al P reval e nce Su rveys

Cross-sectional surveys may be concerned with: the presence of disorders, such as diseases

or disabilities within a population; measures of health, such as blood pressure, fitness or diet;

and factors associated with health and disease, such as socioeconomic status and

demographic data (Abramson 1991).

Cross-sectional prevalence surveys are most often used to learn about risk lactors lor

diseases ol slow onset and long duration (Kelsey 1986). Other uses include: the promotion

of heatth in a specific group or population studied;the enhanced clinical care of individuals;

and the provision of new knowledge, with inferences that could be applied beyond the

population studied (Abramson 1 991 ).

62 Chapter 1

Cross-sectional surveys have two principle advantages; they are relatively cheap cornpared to suÞ other types of surveys and the sample is taken from the general population, not a small

group (e.g. hospital patients), and therefore the results are generalizable. There are three

major difficutties (Ketsey 1986). Firstly, it may be difficult to separate cause and effect.

Secondly, a series of prevalent cases will have a higher proportion of cases with diseases of

long duration than a series of incident cases, (That is, people who either recover or die lrom a

disease quickly have less ol a chance of being included in the disease group). lf there are

differences between long or short term sufferers in certain characteristics, the association of

these characteristics with exposure and incidence may be obscured' Thirdly, there are

possible problems with definition, for instance, should a treated person be counted or should

a person in remission not be counted.

Weighing up the advantages and disadvantages of cross-sectional prevalence surveys it

appears that they are cost etlective and useful for both descriptive and analytical work,

including hypothesis testing (Abramson 1991). They are also often used to provide base-line

information for more detailed planning of curative and preventive work (Minassian 1988).

There are, however, several issues that must be addressed if comparisons are to be made

between cross-sectional surveys.

Compariso n Between SurueYs

Although there are ditficulties in comparing results from ditferent eye heatth surveys, an

appreciation of change over time is greatty enhanced by the utilization of quantitative rather

than qualitative data, even if they are generated by several cross-sectional surveys rather than

cohort or tongitudinal studies. Anecdotat evidence submitted as fact can be misleading rather

than informative and can also be subject to bias and misrepresentat¡on. Quantitative data can

also be interpreted seleclively, but if the methodology, assumptions and analysis are clearly

stated, conclusions can be checked and challenged.

Several factors may influence the interpretation of differences between ocular surveys

(Meredith et al. 1989). For example, there may be differences ¡n the ability of workers to

63 Chapter 1 identify trachoma. This may be due to differences in the levelof experience, or lt may refled levels of expertise, personal mot¡vation and expectations, environmental conditions, facilities for testing and fatigue.

The classification of trachoma is another factor which may influence the results of comparative research. Over many years the grading system for trachoma has changed, reflecting both the evolution in the understanding of the disease caused by Chlamydia trachomafis and a realization that a complicated system added little to the interpretation and application of the results. A simplified system has the added benefits both of being usefulfor non'specialized health workers who could be instructed in its use, and of being interpretable by non-medical personnel.

When comparing surveys which have used ditferent grading schemes ditficulties adse' For inslance, although the terminology may be similar, the definitions for disease may be diflerent and could result in significant errors. For example, in the 1985 National Trachoma and Eye

Health program Review (T.E.H.P. Report 1985), a new diagnostic criteria for trachoma was devised by the investigators because it was felt that the old system failed to pick up genuine cases of follicular trachoma, and whilst the classification of limbal lollicles could have been

used to diagnose follicular trachoma, it could not be used to assess the severity of follicular trachoma. When the investigators examined the 1985 results and those ¡n the 1980 R.A.C.O.

Report, they found that by applying the new diagnostic criteria to the 1980 results, a 167o

increase in the follicular trachoma rate occurred, with a22!" ¡ncrease ¡n the prevalence of

severe follicular trachoma. A change ¡n the diagnostic criteria had resulted in a significant

increase in the 'f 980 trachoma prevalence rates.

When comparing two surveys, factors that influence selection of a participant in a survey may

be distributed differentially between the survey populations leading to a response bias. lt

would be difficult to predict and correcl response bias if ¡t had occurred in a past survey and

had not been documented. Differences between surveys could then be due to ditferences

in response rather than actual differences in prevalence rates. A practical example of this

64 Chapter 1 would be a survey of the prevalence of blindness in a community, in which an eye health survey might identily 90% of all blind people. A second survey two years later may not attract the same ¡nterest as the first, with 50% ol all blind people knowing that being seen again would not help them. The results of the second survey would falsely show a fall in the prevalence of blindness.

Differences in the age structure of two survey populations may also lead to errors when estimates of disease prevalence are compared. This may occur for two reasons. Firstly, with trachoma, active infection is found predominantly in the young, a survey population which has a higher proportion of young people will have an apparently higher crude rate of infection due giving solety to the higher proportion of infected young people. This may be overcome by age-specific rates or by age adjustment, in which a single summary rate is calculated that takes into account any dilference in the populat¡on structure, a procedure called age- standardization. Secondly, an increase in population density may increase the prevalence of trachoma. When two populations are compared each having the same population distribution, an equivalent result may be accepted as indicating equivalent socioeconomic or infrastructure development, the true result possibly being confounded by the different populat¡on

densities. Again, procedures should be implemented to adjust for the confounding factor, or

the d¡ÍÍ¡culties with the comparison acknowledged.

Differences in the way the results of a survey are reported may make comparisons with other

surveys difficult or impossible. For instance, if the source population is not delined or the

methodology not descfibed fully, it may be inaccurate to compare the survey results because

differences may be due to differences in survey design rather lhan disease prevalence.

Surveys can be used for surveillance, which perm¡ts the identification of changes in health

status. For example, changes in risk fac{ors for cardiovascular disease have been monitored

by repeated cross-sectional surveys in the Multinational Monitoring of Trends and

Determinants in Card¡ovascular Disease (Monica Project) sponsored by the W.H.O. (Pisa

1989). Case registers may be used for the purpose as well (Thompson 1989)' When

65 Chapter t changes in the prevalence of a disease is noted from repeated cross-sectional surveys care must be taken in ¡nterpretation. For example, incident data should be used lo ¡ndicate changes in the risk of developing a disease because changes in the demographic characteristics of the population e.g. an ageing population or migration of aflected people away from the survey area, may explain changes in prevalence data. Also, as explained above changes in the method of case identification, survey methodology, or disease definitions may

allatfect the results and the¡r subsequent interpretat¡on (Abramson 1991).

lt has been shown that when performing repeated cross-sectional surveys there are

advantages in using separate sample (between surveys) designs, these are (Salonen et al.

1996): 1. differential re-examination response bias is avoided (individuals do not change their

response because they have been examined before) 2. the representiveness of the sample

remains the same. Similarly re-examination of the same cohort has advantages, especially if

the communities are small, these advantages include : 1. the ability to adjust for base'line

differences between communities on an individual level 2. sample sizes may be smaller 3.

there are increased opportunities to explain interactions between base-line characteristics

and risk faclor changes. The major disadvantages are that as the cohort ages and dropouts

occur, its representiveness of the total community is lost, and the survey may affect the

behaviour of the ParticiPants.

Lasily, the evaluat¡on of intervention programs for diseases that have epidemics (e.9.

measles) and therefore have swings in incidence that are naturally high and low, may not be

uselulty monitored using repeated prevalencg surveys, because it would be diflicult to

attribute changes to the program or to natural flucluation (Cutls 1988).

66 Chapter 1

Summary

There are several methodological problems associated with eye disease epidemiology.

These include: sample size considerations, standardization of disease classilications, the

minimisation of errors due to ophthalmic measurement (e.9. in visual acuity testing), the

reproducibility of observations (inter-observer and intra-observer error), and the selection of

an appropriate survey or study design. When surveys are compared these issues must be

considered; also, when the study is a cross-sectional survey several other factors must be

considered. Chiefly, differential participation, the representativeness of the samples and the

naturalfluctuations of the disease being monitored.

67 Chapter 1

1 .5 Concluslon

It is clear from the loregoing chapter that a review of eye health within a community is reliant on the accurate assessment of visual acuity. A sample survey can be the most praclical method of achieving this aim. However, because sample surveys are cross-sectional, comparisons with previous literature can be difficult. Comparisons with international data are also problematical and therefore may only be usefulfor illustrative purposes.

The major causes of blindness in the AboriginalAustralian are trachoma, cataract and trauma.

Trachoma is a readily preventable cause of blindness and lends itself to intervenlion at primary, secondary and tertiary levels of treatment. The prevalence and severity ol trachoma reftects the environmental and socioeconomic conditions prevailing in Aboriginal communities. An assessment of the prevalence of trachoma could therefore yield information about the general health status of a community and the level of its socioeconomic development. Additionally, if comparisons could be made with previous evaluations of eye health, conclusions about progress in a community might be possible.

Chapter one also ind¡cates that there are several methodological problems when assessing eye disease in a community. Any study has to define the number ol individuals to be examined and make a choice between counting subjects or eyes. Standardization of disease definitions and ophthatmic measures is important, especially if comparisons are to be made with previous surveys. Also, observer variation needs to be acknowledged, with the

measurement of observer variation is best made using the Kappa statistic.

Although it appears that there are disadvantages w¡th cross-seclional surveys when they are

used to monitor change over lime, if adjustments are made for changes in disease

classification, population age structure, survey design, result reporting and cohort

representation, then for many diseases there are many advantages in using this type of study

design.

lf a sample survey was to be undertaken the lilerature ¡nd¡cates that the lollowing ¡ntervent¡ons

would be of benefit to the community being surveyed: case identification followed by

individual treatment (lndividuals with trichiasis could be oflered epilation or surgical

68 Chapter 1

¡ntervention, those w¡th cataracts, surgery in a major centre); increasing the awareness in both the community and medical services of preventive and treatment measures for eye diseaSe; and specilic trainíng of community heatth workers in eye care.

69 Chapter 2

Chapter 2. Blindness and Trachoma ln Australla wlth Reference to the Anangu Pitlantlatiara of South Australla

70 Chapter 2

2.1 Introductlon ln this chapter the causes of blindness in Australia are reviewed. The literature relating to the

Aboriginal population is more comprehensive than that available for the white population' mainly because of the contribution the N.T.E.H.P. survey 1976-1979. lt is important to understand the differences in eye health between white and Aboriginal Australians because the equitable distribution of health resources is dependent on the accurate assessment of need

Trachoma is an important cause of blindness in the Aboriginal population and has been historically important to their eye health. Furthermore, its origin in Australia has been subiect to speculation for many years, with no firm epidemiological basis for definitively ascribing its

¡ntroduction to the coming of Europeans in the 18th and 1gth centuries. A review of tho historical evidence and epidemiological data allows further conclusions about its orígins in

Australia to be made.

The literature review is also used to construct a picture of Aboriginal eye health in South

surveys Australia f rom the earliest recordings by explorers in the 19th century through several

of Aboriginat people in the north-west of South Australia from the 1960s to the 1980s. No

other group in Australia has had its eye health recorded and published as much as the

Anangu p¡tjantjatjara of S.A.. lt provided a unique opportunity to make descriptive

comparisons between surveys over time and to draw conclusions about the development of

trachoma in their communities.

71 Chapter 2

2.2 Bllndness ln Australla

probably more wo¡t has been done to establish the prevalence of blindness in the Australian

Aboriginal population than arnongst the white population. Early reference to Aboriginalocular

disease and blindness occurs in anthropological review articles by Cleland (1928) and later

packer (1961). The first major ophthalmological surveys began in the 1940s with Flynn (1957)

(1954 1954 in Central Australia. These were lollowed by work in Western Australia by Mann a' in the 1960s by b, 1957 a ) in fhe 1950s and several smaller surveys in S.A. and the N'T.

Moore et al.(1965), Mann and Rountree (1968) and several unpublished workers. ln the Health 1970s the most extensive survey carried out was by the National Trachoma and Eye

program (N.T.E.H.p.), whose Chief Ophthalnnlogist and director was Professor Fred Hollows.

Beginning in 1976, in the arid areas of central Australia, 62,116 Aboriginal and 38,616 non-

Aboriginal people were examined (R.A.C'O. Report 1980).

presentation of the results for some of these surveys allows qualitative comparisons to be

made. For instance comparison of the causes of blindness in the whits and Aboriginal

population highlights that there will be differences in the strategies for prevention in each

group. The estimates of blindness prevalence lor each population demonstrate the order of

magnitude separating the two populations in terms of their eye health. Tables 2.1.1, 2.1.2

and 2.1.3 summarise the results of some of these surveys ol Aboriginal and non-Aboriginal

people.

72 Chapter z

Table 2.1.1: Msual system abnormalilies in the blind by racial group

(rates per thousand) R.A.C.O. Report 1980

Rates Per 1000 blind Site Aborigines Non-Aborigines Rate Rank Rate Rank

SIte lens 641 1 516 1 cornea 555 2 83 4 eyeball 149 3 33 11 conjunctiva 118 4 50 6 eyelids 87 5 33 10 iris, ciliary 47 b 66 5 body refraclion 46 7 183 3 slrabismus 24 8 50 7 retina and 18 9 250 2 choroid optic nerve 17 10 lacrimalsystem 16 11 vitreous I 17 sclera 1 18 Cause glaucoma 16 12 33 visual 9 13 16 13 disturbances general 5 'l 4 50 I disorders congenital 4 15 16 14 abnormal¡ties neoplasm 3 16 16 12

The table shows that for many blind individuals ons or more sites or causes of blindness were

pathology recorded by the examining ophthalmologist. This is not unusual, because muttiple

often exists, and it is ditficutt to attdbute the reason for blindness lo one cause. Therefore the

numefators of the rates given may add up to more than a thousand.

73 Chapter 2

(e.9. were the The table demonstrates that for both populations diseases of the lens cataract), the next leading cause of blindness. Corneal (e.g. trachoma) and "eyeball" d¡seass were then two most important diseases for Aboriginal people, but for non-Aboriginal people diseases of the retina or choroid and refractive problems follow lens disease. This is an important dilference, for while corneal and eyeball disease may be preventable, retinal and choroidal disease (e.g. macular degeneration) is more a function of ageing and not preventable with current knowledge.

The data lor the Eastern Goldfields survey of Aboriginal blindness (Table 2.1.21 were derived from a smaller sampte (2,209 vs 100,732 for the R.A.C.O. Report of 1980), and Mann choose

to deline the causes of blindness explicitly. She also choose lo provide information about

both eyes. Twenty three individuals had a single cause of blindness listed (e'9. cataract), but

five individuals were listed as having two causes (e.g. trachoma and cataract or cataract arìd

injury). whilst it is often necessary lo narrow the cause of blindness to one disease, it can be

useful and more accurate to list the cause of btindness in each eye, preserving the detail of

the data.

Table 2.1.2: Causes of Blindness in the Eastem Goldfields

(based on2203 people surveyed) (Mann 1957 a )

Cause of Blindness No. of Cases Rate per 1000 ¡n those surveyed

Trachoma 16 7.3 Catarad 4 1.8 Cataract and injury 3 1.4 Trachoma and cataract 2 0.9 Ophthalmia neonatorum, corneal scars and bilateral phthisis bulbi 3 1.4

Total 28 12.8

74 Chapter 2

next most Mann found trachoma to be the principal cause of blindness, with cataract being the important cause. The survey was small; however lhe data, in particular that relating to trachoma and the high prevalence of blindness, was of great interest at the time'

it The data in Table 2.1.3 were a sub-set of that previously discussed (Table 2.1.1l,' However, provided gave specific disease categories rather than sites of visuat system abnormalities, and

had information for both eyes. lt was a larger sample than that for the Eastern Goldfields, and

group rather than several disease categories. The results tend to mirror those of its larger data was due those ol Mann's work. Taylor (1980 a ) found that almost half of Aboriginalblindness

to cataract, one third to corneal scarring and one twelfth to phthisis or enucleation. Only one people twelfth was attributable to ret¡nal disease. He concluded that blindness in Aboriginal

was environmentally related and therefore preventable. Trachoma contributed to only 18.1%

(9/55) of all binocular blindness, a finding consistent with the N.T.E.H'P. data, but less than geographical that obtained by Mann. Mann's and Taylor's work were completed in ditferent

locations, and this may account lor the observed differences. The R'A.C'O. Report found

considerable variation between the 13 zones ¡t had chosen, which reflected both cl¡matic and

socioeconomic conditions (R.A.C.O' Report 1980).

75 Chapter 2

Table 2.1.3: Causes of Blindness ln 138 Aborigines Over The Age

Of 60 Years; n = 339 (Taylor 1980 a )

Binocular Monocular No. No.

Trachoma 7 5 * with cataract 1 with enucleation' 1 Comealscars I 15 * with cataract 2 with phthisis' 1 Senile cataract 32 25 traumatic' 1 with pt.' 1 with enucleation' 2 Traumatic cataract 5 Climatic droplet keratopathy 4 3 with comeal scar' 1 Bilateral phthisis I 12 Other 3 I

All causes 65 73

. tn these cases of binocular blindness, one eye was atfected by lhe cond¡t¡on listed in the major entry;the other eye was affected by the condition ¡n the sub-entry.

ln contrast, very little literature appears for the urban white population. Anderson (1939)

reported the prevalence of blindness and its causes among private pract¡ce patients in

Melbourne. He conducted an analysis of the medical records of 12,240 ophthalmology

patients. The definition of blindness used was vision of less than 6/36, or a field loss reduced

to within 1Oo ¡n thê other. The survey was acknowledged as incomplete as ¡t did not sample

from the whole community. He found binocular blindness in 2.2"/" of the sample and

monocular blindness in 5.5%. CI more interest were the listed causes of blindness. ln order

of importance they were: glaucoma; congenital and hereditary diseases, myopia, vascular

lesions, syphilis, diabetes, focal sepsis and trachoma. The results reflect the type of patient

seen in privale pracrtice, the treatment available at that time (e.9. for glaucoma or syphilis), and

76 Chapter 2 the exclusion of people with cataract, as they were classified as "amenable to treatment"

(Anderson 1939 P680). ln his discussion of the results, Anderson concluded that 50% of blindness was preventable with the then current knowledge, and a further 30% possibly preventable in the future. He was concerned at the implication of the results for monocular blindness, part¡culafy the greater risk of complete blindness and the impairment of efficiency if the other eye was not good. The presentat¡on of the results for blindness from trauma was comprehensive, and gave a detailed picture of events and objects that had caused monocular or binocular blindness. Motoring accidents, chopping wood, shooting and hammering stoel being the predominate causes. Although the study was acknowledged as being limited, it was hoped that the work would encourage a more systematic invest¡gat¡on of blindness in Australia.

During the Second World War (1939-1945) the results of a blindness survey conducted in

Tasmania were published, with a final review in 1948 (Hamilton 1950). lt presented the results

of a private practice audit in Hobart, with a comparison to similar data collected in 1938. The

work was criticized by ophthalmologists of the day for not being represenlalive of the total

incidence of blindness in Tasmania, as it onty concerned private pat¡ents in Hobart, further,

there were potential problems with double counting of patients between practices and the

totat number of blind pat¡ents seen (106), which was thought to be small. The compadson

with earlier work was seen as ¡nteresting; however, other explanations for the observed

changes were not explored. For instance, although the decline in blindness from diabetes

may have been due to the introduction of insulin, other factors such as the referral of diabetíc

patients to Dr. Hamilton may have been important (Hamilton 1950 : Discussion).

ln 1949 Redmond presented the find¡ngs of a review, in which the records of 10,000 pat¡ents

(from a ten year pedod) were examined, and the causes of bilateral and unilateral blindness

described. The study was not presented as a paper on blindness, but as a stimulus for other

77 Chapter 2 ophthalmologists and the government to consider both the lack of data on blindness in

Australia and the prevention ol unilateralblindness (Redmond 1949).

Following on from this early work, the Ophthalmological Society of Australia proposed to

evaluate data from the compulsory registration of blindness. A form, the SA 47, was designed

lor the certification of people seeking the invalid pension. lt was introduced in an effort to

obtain accurate statist¡cal information on the causes of blindness. A survey was run for five

years from 1953 to 1958 inclusive (Yates 1963). Although the data was subiect to bias, it was

useful as a guide to the causes of blindness in Australia. The principle causes were: cataract

17.7go/o, glaucoma 16.02Vo, diabetes 6.89%, and all other causes 11.38%. Trachoma was

found to cause 2.'t7% of all blindness (it was not clear whether Aboriginal people were

counted in the data). Yates compared the prevalence rates with those obtained by Sorsby in

England, and Mann in the Kimberleys and Eastern Goldfields. He found that the causes of

blindness were similar to those found by Sorsby in England. However, when he reviewed

Mann's results from the Kimberley and Eastern Goldfields ¡t was apparent that none of blind

individuals f rom these areas appeared on the SA47 forms.

A later survey estimated major causes and prevalence of blindness based on four sources

(Banks and Hutton 1981):

1. Records of the Department of Social Seotdty.

2. The Nat¡onalTrachoma and Eye Health Pogram.

3. Australian Bureau of Statistics.

4. The RoyalBlind SocietY.

Aboriginals were not included and methodological probtems were found with each source of

information. The estimated prevalence of binocular blindness was of the order of 7 to 10 per

lOOO (0.14-0.19%). of these, onty half were receiving a blindness pension. Banks

concluded that the although an absolute value lor blindness prevalence could not be

obtained from pension data, the % contribution of any cause could be readily available with

78 Chapter 2

coding for cooperation from the Department of Social Security, and by adopting interacfive listed by diagnosis in the Department's records. The causes of blindness so obtained were site of pathology, making comparisons diflicult. ln order they were: macular degeneration other 37.zyo, anterior segment 28.8o/o, neurological (including optic nerve) 11.zyo, and genetic 1g.S%. Acquired cases made up 67% ol all blindness, with congenital 10% and

15.5%. He concluded that aCCUrate information could only be obtained from extensive population based surveys.

The presentation of these results for blindness in Australia demonstrates the ditficulty ol obtaining accurato prevalence data and while the causes ol blindness wero more easily particularly identilied, the percentage contribution ol each was subiect to ascerta¡nment bias, for data obtained from private practice or social security records. The former eliminates those who cannot atford private health insurance, and the latter those ineligible or unwilling to obtain the blindness Pension.

2.9 Orlglns of Trachoma ln Australla

The ancestors of today's Aboriginal Australians crossed a land bridge between Asia and

Australia somewhere between 60,000-40,000 years ago. They migrated south through a

levels, land which was wetter and greener than it is today (Abbie 1976 pxviii ). Wth tising sea

they were then effectively isolated until the coming of Europeans in the 17th and 18th

centuries. From what is known of the Aboriginal lifestyle prior to white settlement, it appears

they were nomadic or semi-nomadic, hunting and gathering in small groups (Bemdt and

Berndt 19771. lf trachoma had been endemic in Australia it would have necessarily survived

by 40,OOO years of changing environments, as Australia became drier, and mainta¡ned itself

carriage and transmission in small groups of scattered people, who it appears had few other

infectious diseases, and were physically lit and healthy. Given these constrainls, and the fad poor it that trachoma is difiiqJlt to catch and only thrives in crowded commun¡ties with hygiene,

79 Chapter 2 appears that the presence of trachoma as a blinding disease did not occur untilthe advent of the socio-cuttural and bacteriological changes associated with the coming of white man.

There were ethno-historical accounts of eye disease in Aboriginals from the 19th century

(sturt 1g33 , Eyre 1854, Stirling 1894). AlsO, a number of anthropologists have speculated that trachoma was present prior to white settlement (Abbie 1960, Barnes 1970), and some authors on Aboriginal life have used slatements attributed to explorers to suggest that trachoma may have been present priorto white settlement (Cleland 1928).

The lirst known European explorer to visit Australia was William Dampier (1688). He wrote ' ..' the m¡serablest people in the world, their eyelids are always half closed to keep the fl¡es out of their eyes, they being so troublesome here that no fanning will keep them from coming to

one's face so that from infancy, being thus annoyed with these insects, they do never open

their eyes as other people, and therefore they cannot see far unless they hold up their heads,

as if they were looking at somewhat over them'. The Professor of Pathology at the University

of Adelaide speculated that the above passage referred to trachoma (Cleland 1928). lt may,

however, have been another form of eye disease (R.A.C.O. Report 1980). ln contrast, during

the voyage of the "Endeavouf in 1776, the English explorer Captain James Cook (1788), did

not observe any sign of disease among the Aborig¡nals encountered. Further corroborating

evidence from early settlement is currently not available, but it appears that there is no

indication in the traditions or folk lore of Australian Aboriginal people, as lo the origin of

trachoma (Mann 1957 b ). Such information might be expected il it had been present for

40,000 years.

With the colonization of Australia aller 177ï,trachoma was acknowledged as ocolrring in both

the wh¡te and Aboriginal population (Cleland 1928). Confusion has arisen between endemic

trachoma and epidemics of conjunctivitis, with the Australian term'Sandy Blight" once

thought to describe trachoma (Editorial 1972), now being recognised as a non-trachomatis

conjunctivitis (Hollows 1989). Whateverthe terminology, severalauthors have speculated on

the origin of trachoma. ln 1904 a European author, in a review of trachoma throughout the

world, supported the contention of its European introduction (Boldt 1904), " lt is probable

80 Chapter 2 thal the disease was introduced by the convicts who were transported". Boldt observed that trachoma was prevalent amongst the inmates of prisons and other institutions in England at the beginning ol the 19th century, from which the early convicts transported to Australia would have been drawn.

Mann, an ophthalmologist from Western Australia, conducted several trachoma and eye

health surveys in Australia and Papua New Guinea during the 1950s (Mann 1954 a, Mann

19S4 b, Mann 1957 b, Mann 1967). She found strong epidemiological evidence lrom the

distribution of trachoma in New Guinea to support the hypothesis of its recent introduction

there. By contrast, in Australia she found trachoma to be widespread in allthe areas surveyed.

Although acknowledging that it may have been endemic, she speculated that Western

Europeans (setilers, convicts or goldminers), Japanese or Malay pearl fishers, Afghan camel

drivers, imported Chinese labourers, natives lrom other Pacific lslands, and Central and

Eastern European m¡grants, could have introduced trachoma in more recent times (Mann

1s57 b ).

lf trachoma was introduced by the early explorers and settlers, its transfer and early disastrous

effects on the Aboriginal population went unnoticed. ln addition, the socio'cultural

disintegration seen when large groups of Aboriginal people were forced into crowded

communities probably enabled the disease to flourish (R.A.C.O. Report 1980).

An opportunity arose in the 1960s to examine a group of Aboriginals from W'A. who had no

previous associat¡on with Europeans (Elphinstone 1971). He reported that ol 88 individuals,

half had signs of active or healed trachoma, with one case of trichias¡s and tvro cases of comeal

opacity due to trachoma. lt could be concluded that this finding provided evidence ol

trachoma's ancient origins in Australía. However, it was known that this group had had some

contact with Aboriginal people from surrounding areas. These Aboriginals were infected with

trachoma. lt could be argued that the lower prevalence of trachoma, and its severe

complications (of trichiasis and comeal opacity), when compared with results from the adiacent

Warburton area (from Mann's 1957 survey), indicated a more recent introduction into this

8f Chapter 2 group, with insutficient time for the disease to infect all susceptible individuals, or run a chronic course, in order to produce the blinding complications so characteristic of trachoma in endemic areas.

The only other evidence concerning the origins of trachoma in Australia, comes from the 1,000 to examination of the skeletal remains of prehistoric Aborigines (Webb 1990). Skulls

14,OOO years old, were examined for a small round hole ¡n lhs orb¡talplate ol the frontalbone.

Webb speculated that a chronic infection of the lacrimalglands was the most likely cause and

and that trachoma was the most probable agent. The lesion was positively conelated with age was distributed in trachomatous regions. However, males were atfected rnore than females (a

reverse of the epidemiology today), and practising ophthalmolog¡sts doubt the association

between the lesion and aclive trachoma, they think it may have been due to post mortlid

trauma. (Personal communication Dr. J. Crompton 1991, Adelaide, South AJstral¡a)

2.4 Past Surveys ln South Australla

The first observations of the eye health of Aboriginal people in South Australia (S.4.) were

made by explorers from the 1880s to the 1920s. They recounted anecdotal ev¡dence of eye

disease. For instance, the use by Adelaide Aborigínal people of mast¡cated gum leaves for

sore eyes caused by "Sandy Blight" was described by Stephens in 1889 (Cleland 1928).

Ophthalmia was stated to be prevalent amongst the natives in Central Australia encountered

on the Horn Expedition (Stirling 1894). ln 1916 several old Abodginals were met ¡n the north

of South Austratia and found to be blind (Waite 1916). The Professor of Pathology at the

University of Adelaide later ascribed this to " ...the results of trachoma or sandy blight. "

(Cleland 1928 P219).

Trachoma was also noted to occur in the white population. ln a review ol trachoma thoughout

the world Boldt (1g04), reported on the lindings of a Dr Bennett. He was a generalpraclitioner

for 8 years in a small town called Crystal Brook, which is located in the mid-north of South

82 Chapter 2

drawn Australia, an area known for its farming and low rainfall. An imaginary line had been Known as across the state of South Australia, delineating arable from non-arable land' the line Goyde/s Line, it reflected lhe average rainfall across the state, with the area north ol produce crops receiving less than 200mm of rain per year. Enough rain fell south of the line to consistenily. Dr Bennett found trachoma to be twenty limes more frequent north of Goyde/s had adequate line than south of it, despite the fact that the farmers were otherwise heatthy, diets and lived in spacious houses.

eye disease ln 1g2g, the professor of Pathology at the university of Adelaide reported on the glaucoma of Aborigines. ophthalmia, gonorrhoeal conjunctivitis, corneal opacities, trachoma, 19th ard and senile cataract were discussed with reference to previous observations from the early 20th centuries (cleland 1928). A Dr. Basedow, was acknowledged on several in occasions, for his work in the N.E. of South Australia during three medical reliel expeditions

1g1g-1g20. The article was descfiptive, and gave no indication of the prevalence or picture of incidence of disease. While of historical ¡nterest, it may or may not be an accurate

Aboriginal eye health at the turn of the century.

óoctor Basedow in 1932 contributed furlher information on the "Diseases of the Australian the Aborigines. when he wrote as the former Chiel Medical lnspector and Chief Proteclor of done to Aborigines of the Northern Territory. He lamented the lact that more had not been

document the diseases of the "pure blooded" Aboriginal Australians, desp¡te ample

opportun¡ty by medical officers ¡n the past. He also presented personal observations of

Aboriginal customs, behaviour and health beliefs. He emphasised the influence of traditional

Aboriginal ideas of health and sickness on western diagnosis and treatment of disease

(Basedow 1g32). He was aggrieved, that despite suggestions, "no systemat¡c lreatment

whatsoever has ever been inaugurated in Australia for checking the alarming spread of for trachoma among the natives. To placate the prevailing sentiment at the time he added'if

no other reason, this should have been done long ago in the interest of the white senlers'

(Basedow 1932 Pl32 )

83 Chapter 2

people continued contributions to the limited amounl of knowledge ol diseases in Aboriginal Australia, and in the 1g30s with an expedition to Nepabunna in the Flinders Ranges of south to the'Granites,, 380 miles north-west of Alice springs (Black and cleland 1938).

,,full f the Granites. Onry blood. Aboriginals were examined, nine from Nepabunna and fitty rom eighteen in Corneal opacity was found in four of the nine from Nepabunna, and three of the

in the the Granites. The author commented that corneal opacity was comr¡on and was severe

three, and had reduced substantially the vision in each eye in one woman from Nepabunna. preceding Grey and Cleland (1933) had previously considered that the corneal ulcers the passags opacities were due to injury from three sources: fights, mulga twigs or other during of the through dense scrub, and fire or fire-sticks, especially in the case of children' Some

cases from Nepabunna were thought to be of gonococcal origin. Three subjects from the of Granites were thought to have scarring of the upper lids, possibly due to trachoma. Signs

trachoma were not noted in the individuals lrom Nepabunna'

Blindness was not remarked upon except for the woman from Nepabunna with "very liüle

vision in either eye". Vision testing was found to be unsatísfactory because "¡t was difficult to p81 get them to rivet their attention on what you want them to see" (Grey and Cleland 1933 ).

Refract¡on was estimated on six using homatropine and cocaine. Allshowed some degree of

hypermetropia except one, a young women aged22 years who was emmetrop¡c.

both Atthough the N.W. areas of South Australia were not visited, the Abodginalpeople in the

N.W. of S.A. and around Nepabunna may have been living under simildr conditions and it

could be expected that the pattern of trachoma would be the same.

Flynn conducted several eye surveys in the 1940s. These were the first modern

assessments of eye health amongst the Aboriginal population of central Austral¡a. He

commented in 1942: "At this time it appeared to be the generally held contention that little or in no aclive trachoma existed in Australia'(Flynn 1957 p 269). When he conducted a survey

1944 he found severe disease in the area around Alice Springs, with 90% of those examined

infected with trachoma, and wilh7"/o blind in one or both eyes (Flynn 1957)'

84 Chapter 2

During the 1gsos, Mann conducted Aboriginal eye health surveys in the Kimberley District of from Western Austratia and in the Gotdfields area (Mann 1954 a, 1954 b, 1957 a). The resuÌts population these surveys showed that trachoma was a major problem in the Abodginal and that a large different¡al existed between the eye health of Europeans and Aboriginals'

Treatment programs did not reduce the prevalence ol trachoma appreciably, and Mann living concluded ', ... lit¡e impression will be made on the trachoma rate unless the standard of pl020 of the infected groups can be substantially raised (Mann and Rountree 1968 ).

part South ln the mid 1960s Mann conducled a survey in the Pilbara and the north-western of

Australia (Mann and Rountree 1968). These large regions were chosen because of tho

paucity of white inhabilants, and the fact that traditional tribal life conlinued for the Abodginal

population.

ln this report no details were given of the methodology involved in the survey, but a given descriptive account of the people, the¡r customs and living conditions were with

summary tables of the resutts. Two communities were visited. At Ernabella, 377 Iull blood"

Aboriginal individuals were examined from an estimated 400-500 members of the tribe. At

lndulkana which was litile more than a temporary camp with no amen¡t¡es, with buildings and

nursing services some miles away, 118 individuals were examined.

The results of the Emabella survey indicated a high prevalence of trachoma, 907o overall, with

53% of the under-2o age-group having active trachoma (the under-2o age'group comprised

50% of the total population). Visual aority was tested on only 77 individuals, 61 of whom had

6/6 vision or better. Three elderly people with lens changes read 6/36. lt was of note that only

one person over 60 years of age was 'economicatly blind", and only five showed visual population disability due to trachoma. However, these data may be imprecise, only 20,/"of the

undenrent visual acuity tests and it is unclear how the blind or visually impaired individuals

were identilied. certainly blindness in older individuals was not a maior feature of the survey in

Ernabe¡a. ln contrast the survey in lndulkana, although examining only 118 Aboriginals'

found trachoma in 84To, with a "patlern of distribution which was similar to other surveys of

85 Chapter 2

W.A. and S.4." Mann and Rountree commented that most had healed by middle-life with rþ impairment of vision, the only sign being scarring of the tarsal plate, Herberts pits and mild pannus. Eight persons had corneal opacity with five being 'economically blind' as a result.

No other causes of blindness were mentioned. Tables 2.4.1 and 2.4'2 delail Mann's and

Rountree's results for trachoma status and eye disease.

Table 2.4.1: Ernabella and tndulkana - Trachoma Stalus by Age (Mann and Rountree 1968)

I 2Trachota3 Age-group Total No Traclpma Trachoma Trachoma Tracrpma4 examined r-r (A) ¡!r(A) lv (A) rv(c) & rv(D)

<1 19 12 7 0 0 0

1-4 83 20 56 6 1 0

5-9 68 2 22 39 5 0

10-19 78 1 4 48 25 0

20-59 190 15 0 45 126 4

60+ 57 f 0 3 44 9

Total 495 51 89 141 201 13

% of total 1O .3o/o 17.9% 28.4o/o 40.s% 2.80h

Mac Callam's Stages of Trachorna (Mac Callam 1931)

1. Tradtoma l-ll(A): Folliq¡lar Trachoma

2. Trachoma lll(A):Where cicatrization has commene,ed

3. Trachoma lV(A): Where cicatrization is complete and healed with good vision

4. Trachoma lV(C) and lV(D): Where cicatdzation is conplete and either healed with impaired sight or is economically blind

86 Chapter z

:|able 2.4.2: Distribution of CIher Eye Diseases in Ernabella and lndulkana (Mann and Rountree 1968)

Disease Emabella lndulkana

Perlorating injuries 18 7

Non-perfo rating with inlection 7 0

Lens changes (cataract) 10 1 (nuclear sderosis) 6 0

Eye excised 2

Ruptured sphincter 1

Totalexamined 44 I

Mann and Rountree summarised their findings as showing that trachoma was the most

important eye disease throughout the areas. Perforating injuries to the eye Írom trauma were

more commonly found in Abodginalthan non-Abodginalpeople, and despite the high number

of perforating injuñes, no sympathetic ophthalmit¡s was found.

Several limitations atfected the accuracy of the survey, principatly the representativeness of

the sample. The two major settlements were visited; however, ¡t was not known how many

individuals existed away from these areas, leading nomadic lives. A total of 400-500 may be a

reasonable estimate by locals familiar with the Pitjantjatjara. However, with peopls ranging up

to 300 km from the poputation centres, any estimate was likely to have been a guess'

ln South Australia during the 1960s trachoma was noted to be'extraordinañly rare'in the

white population. Little was known about its prevalence in the Abodginal population ¡n the

state (Mann's work had not been published). Atter isolated reports of active trachoma in

87 Chapter 2

Adelaide, and a meeting of interested groups, it was decided to perform a survey of trachoma in urban and ruralpopulations (Moore et a1.1965).

The urban white samples in Adelaide and Salisbury were highly selected. They were either

patients of ophthalmologists in whom follicular conjunctivitis had been diagnosed, or, in the

latter township, ex-servicemen, many of whom had served abroad. Musgrave Park (Amata)

and Mt. Davies (pipalyatjara and Kalka) ¡n the far N.W. of S.A., and Cooper Pedy were selected

as rural localities. The survey was conducled in 1963-64.

All the rural areas were visited again in 1976 -1979 N.T.E.H.P. survey, with Mt Davies and

Musgrave park included in the 1989/90 survey. The results for these two communities (for

1963) are summarised in Table 2.4.3. Trachoma was found in 96% of the 163 individuals

examined¡nthe1965survey. CornealopacitieswerepresentinlSsubjects(11.0%),with13

(72.2%l of these considered to be trachomatous. Three subjeas showed severe visual loss

(i.8%), two being occupationally blind in both eyes (1.2%). Tñchiasis and severe lid deformity

had occurred in two subjeAs. Moore concluded that his findings indicated a high prevalence

of trachoma, as in previous reports in W.A. and the N.T., but with low rates of severe lid

deformity and blindness.

88 Chapter 2

Table 2.4.3: Musgrave Park and Mt Davies (Moore et al'l965)

Age-grouP vs Trachoma Status

1 2 Age-group total No Stage Stage stages stage4 complications (Years) examined Trachoma I I It lv

0-1 10 2 4 4 2-4 10 10 5-9 19 I 11 ; ; ; 10-14 14 3 5 4 1 1 15-19 11 1 1 b 3 20-29 16 1 8 6 1 30-50 62 4 1 9 30 I 51+ 21 3 i 2 13 3

Total 163 2 16 33 42 55 15

1. Stage I -Early stages of pinhead follicles.

2. Stage ll -Folliqrlar trachoma.

3. Stage lll -Where cicatrization has commenced:often non contagious.

4. Stage lV -Where cicatrization is complete: non'contag¡ous

5. Complications - comealopacity and trichiasis.

A study on the effect of improved hygiene and treatmenl with oxytetracycline and systematic

sutphormethoxine was conducted by Hardy (Hardy et al. 1967). Abotiginalschoolchildren at

yalata in South Australia were in¡tially surveyed in March 1965. Coniunctival smeafs were

taken for cytology. tmproved wash¡ng facilities were provided in May 1965 and a second

survey undertaken in July that year. Because of the high prevalence of trachoma, treatment

with oxytetracycline eye drops for 3-4 weeks and sulphormethoxine (one doseÂveek lor 12

weeks) was underlaken untit a third survey was conduc{ed in October 1965.

It was concluded that ¡mproved hygiene produced a significant change towards a normal

cytological picture. Further significant change was produced by treatment w¡th oxytetracycl¡ne

eye drops and oral sulphormethoxine. This study provided evidence from an appropdately

89 Chapter 2 organised intervention that hygiene was linked to trachoma, reinforcing previous evidence that s¡mple primary health care intervention could be significant in decreasing infectious disease. Additional improvements in the rate of trachoma with medical intervenlion

heighlened the need for a two-pronged strategy, to maximize the elfectiveness of any

intervention Program.

A survey of visual acuity and retinalchanges in South Australian Aborigines was carried out in

1975 (Edwards et al. 1976). Four Aboriginal communities were visited and 361 adults

examined. The examination for eye health was perlormed during a general health survsy.

Visual acuity and retinal vascular defects were specifically assessed. Good response rates

were obtained at three of the four communities. Arterio-vascular changes were seen in 26o/o

of a1 subjects but only rarely observed in those from Ernabella(7%1. Severe changes in the

retina such as haemorrhage/exudates and papilloedema associated with hypertension and

diabetes were not seen although some individuals had been diabetic or hypertensive for 15

or more years. Only 64 (18%) had a visual defect (visual acuity of 6/9 or worse), and an

additional 79 (22%) had a similar loss of acuity in one eye only, these were lound mostly in the

non-traditionalcommunity of Konnibba. Causes of poor vision were not reported.

The largest and most comprehensive survey in Australia was undertaken by the Royal

Australian College of Ophthalmologisls with the NationalTrachoma and Eye Health Program

(R.A.C.O. Report 19BO). Starting in South Australia in 1976 and finishing with a rescreening

of the Pilbara region of Western Australia in 1979, approximately 100,000 people were

screened, some of whom were examined on two occasions. For the purposes of the report,

62,116 were considered to be Aboriginal and 38,616 non-Aboriginal. All the states on

mainland Australia were covered, with mainly rural or semi-urban Aboriginal communities being

visited. To assist in the collation and integration of the collected data, rural Australia was

divided into 13 zones, these having been chosen because of a relatìve similarity in

geographic, climatic, industrial and demographic features. Although the population sample

was not random, the investigators thought the sample obtained was representative. Lastly,

90 ChaPter 2

and intra- efforts were made to standardize the examination and reduce both inter-observer observer variation.

The results detailed age-specific prevalence rates by zones, and collectively, for trachoma and

blindness, visual system abnormalities, and for infections of the ear, nose, skin and respiratory

system. Results were cross-tabulated with a selec{ion of environmental variables, such as

climatic conditions (normal relative humidity, annual rainfall, latitude, intensity ol ultra-violet

radiation, annual evaporative and average daily sunshine hours) and hygiene (water access,

sewerage systems, housing, nutritional status). Composite scores were allocated for each

hygiene and climatic variable. The prevalence rates for several disease and visual function

categories were tested for the strength of association with the climate and hygiene variables.

ln the following discussion the N.T.E.H.P. data dealing wÍth the "Red Centre" zone will be

presented and used later for comparison with the 1989/90 survey (Tables 2'4.4 and2-4.5)-

Comprehensive results were presented for th¡rteen zones across Australia in the original

report. The N.W. communities of the Pit¡antiat¡ara were included in the "Red Centre" zone of

the N.T.E.H.p. survey in 1976, an area that represented 16% of the total population forthat

zone. yalata was included in the "Arid Eastern" zone, whích extended ¡nto western N.S.W.

and north into Oueensland and which included communities in Pt. Augusta and Cooper Pedy

(S.A.) Bourke and Wilcannia (N.S.W.), and Longreach (Old.). Yalata contr¡buted only 211 out

of 8671 (2.4%) individuals to this zone.

91 Chapter 2

Table 2.4.4: Age Distribution of Aboriginals in the "Red centre'zone

Age

0-9 10-19 20-29 30-39 40-49 50-59 60+ Total

7997 Number 2944 1 997 929 614 486 335 692

100"/" ôtto 36.8% 25.0% 11.6o/0 7.7"/o 6.1% 4.2o/" 8.7"/o

Table 2.4.5: Rates (per 1000) for visualaotity in the'Red centre" zone

Age in Years 0-9 1 0-19 60+ Allages 0-60+

Good Vlslon 1 Rate/1000 994 987 361 908

Satlsfactory Vlsl on2 rate/1000 ooo 997 s55 952

Poor Vlslon 3 rate/1 000 I 2 220 26

4 Monocular Bllndness rate/1 000 3 5 227 37

5 Blnocular Bllndness

rate/1 000 225 23

(r) Good vlslon: corrected visual acuity ol6112 or better in both eyes

(2) Satlsfactory Vlslon: Corrected visual acuity ol6112 or better in one eye

(3) poor Vlston: Corrected visual acu¡ty ol6t12 ot 6124 or 6/36 in the better eye.

(4) Monocular Bllndness: Corrected visual acuity of no better than 6/60 in the worse eye and vision better than 6/60 ¡n the best eye

(5) Blnocular Bllndness: Corrected visual acuity ¡n the better eye of 6/60 or worse.

92 Chapter 2

Foiliculartrachoma was found in26o/" of the 7982 Aboriginal people surveyed in the'Red peak Centre" zone. !n the O-g age-group the prevalence of follicular trachoma was 50%. The prevalence occurred in 2-3 year olds (66.3%). The lowest prevalence was found in the 18'19 year olds (12.00/"1. Cicatricial trachoma was found in 5149 Aboriginals (65%) overall, and in

95% of Aboriginal people aged 60+ years.

The results demonstrated the poor eye health of Australian Aboriginal people. The report also indicated the impact made by trachoma on the Aboriginal people living in Australia, and the relationship between trachoma and poverty. Several recommendations were made relating to the provision of adequate housing, water and sewerage, and to the luture role ol the N.T.E.H.p. in providing trachoma and ophthalmic services to ruralcommunities.

ln October 1985 the NationalAboriginal and lslander Health Organisation presented a report on Trachoma and Eye Health to the Hon. Clyde Holding, Minister for Aboriginal Atfairs' lts lerms of reference were (i) to report on the current ocular health status of Australian

Aboriginals in a manner comparable with that provided in the R.A.C.O. Report of 1980, (ii) to

assess the effectiveness of the existing anti-trachoma program, and (iii) to promote plans to

effectively deal with the trachoma and eye disease found'

Approximately 2000 Aboriginal people were examined across S.4., the N.T., and northem

W.A. The communities of Amata, Fregon and lndulkana in the north-west of S.A., and Yalata

to the south, were visited.

ln the Anangu P¡tiantjatjara communities the prevalence of follicular trachoma, which had fallen

aler a treatment program in August 1976, had risen by 1995 to levels higher than those

found in the first screening in June 1976. Severe scarring in the up to 19 years age€roup;

however, had shown a drop in prevalence. The reviewing ophthalnnlogist concluded that the

prevalence rate for trachoma was of clinical concern, but formed only a part of a matrix of ill

health. The results lor Yalata were similar but not statistically significant because ol the small

numbers surveyed (T.E.H.P. Report 1985)'

93 Chapter 2

2.5 Dlscusslon

From the foregoing review it appears lhat the high prevalence of blindness in Aboriginal

Australians arises from causes that are currently preventable. The three sets of results (Tables

2.1.1.-2.1.3.) also demonStrate the different ways data may be presented, each revealing

complementary aspects of the data.

The review also dernonslrates the paucity of knowledge about the causes and prevalence of

blindness in the white population in Austral¡a. The ditficulty in obtaining accurate prevalence

data appears to be the principle reason for inaccuracy and population based cross-sectional

surveys appear to be the most accurate way of obtaining information. However, the qualitative

data presented does demonstrate differences between white and Aboriginal populations in

Australia, both in ths prevalence and causes of blindness. Amongst the preventable causes

of blindness in the Aboriginal poputation, trachoma and cataract appear to be the most

important.

It has been speculated, but not proven, that trachoma was endemic in Australia before

European settlement. ln South Australia the earliest recorded observations of trachoma in the

white poputation were reported by Dr. Bennett (Boldt 1904). Anecdotal descriptions of

Aboriginal eye health in the 1880s were used by authors ¡n the 1930s as evidence of

trachoma infection, unpublished observations by Dr Basedow did give some corfirmation that

trachoma existed in the "native" population of South Australia at that time (Cleland 1928). But

it wasnl until the 1930s that an anthropological expedition to Nepabunna in the Flinders

Ranges described in more detail the eye health of some Aboriginal people. No trachoma was

found (Black and Cleland 1938).

ln the 1940s and 1950s Flynn and Mann conducted eye health surveys in Central and

Western AustralÍa respectively. Widespread and severe trachoma was found in both regions.

Onty in the 1960s were the N.W. areas of South Australia visited. Although a high prevalence

of trachoma was lound, few individuals showed visual disabilily because of trachoma (Mann

94 Chapter 2

and Rountree 196g). Mann recorded that nrost of the trachoma had healed by mid-life with only fine tarsalscaning and some Herberts pits and minima! pannus.

prevalence of Similarly in the same area, Moore found a high prevalence of trachoma but a low severe lid delormity and blindness secondary to trachoma (Moore et al' 1968)'

rate of It was not until 1976 that a high prevalence of trachoma accompanied by a high trichiasis and corneat opacity was identified in the N.W. of South Australia (R.A.C.O' Report lgg9). Trachoma was also found to be the major cause of visual disability and blindness. After that another 14 years trachoma was stillthought to be endemic in the AborigÍnalcommunities comprise the A p lands. However, its severity and contribut¡on to blindness, after several years of intervention, was unknown.

To answer these quest¡ons quantitative data needed to be obtained. Although there have been numerous attempts to quantify blindness in Australia it appears only the N.T.E.H.P. surveys have provided comprehensive data adequate for statistical analysis. Chapter three describes a population based prevalence survey used to obtain data in 1989-1990. This data

is then utilized in chapter four to make quantitative comparisons with the N.T.E'H'P. survey

(R.A.C.O. Report 1980) and the N.T.E.H.P. review (T.E.H.P. Report 1985). Methodolog¡cal

problems with the setection of the samples, grading and reporting of results precluded the

use ol other survey information other than for qualitative descriptions.

95 Chapter 3

Ghapter 3. A Survey of Trachoma and Bllndness ln the Anangu Pitlantiatiara: The 1989/90 Survey

96 Chapter 3

3.1 lntroductlon

population To assess the current eye health of the Anangu Pitjantjatjara a cross-secÍional

based prevalence survey was designed.

Although a prevalence survey does not require a hypothesis, it was hoped to make a

comparison between the results obtained, w¡th those of the National Trachoma and Eye

Health program survey of 1976 conducted in the same area. A testable hypothesis was

therefore framed as

"The age-specilic prevalence rates ol blindness and trachoma in the Anangu pitjantjãtjara population of South Australia have shown a stat¡stically significant decline since the N.T'E.H'P. survey of 1976'.

Specific aims and objectives were formulated, including service roles which were to be

important to lhe success of the survey:

prevalence and visualimpairment in 1 . To re-establish base-line data on the of blindness

the Anangu Pitiantjatjara of South Australía.

2. To identify the causes of blindness in this population.

3. To identity trends in the prevalence and population distñbution of eye disease.

4. To assist in the education of community health workers and the people in the

Aboriginal communities about common eye conditions'

S. To identify individuals with blindness and eye disease and facilitate theirtreatment

and review.

6. To analyse and distribute the accumulated data to:

a. TheAboriginalcommunities(AnanguPitiantiatiara).

b. Nganampa Health Council (see section 3'2.3.).

97 Chapter 3

c The South Australian AboriginalTrachoma Program (see section 3.2.3.).

These aims corresponded with the strategies for working with Aboriginal communities, outlined in the report of the National Health Strategy Working Party (1989).

3.2 Methods

3.2.1. Geographlc and cllmatlc Descrlptlon of The study Area

Two distinct areas were visited. Although separated by 1,500 km, they were similar in climate

and culture (Fig. 3.2.1.1 and appendix 6).

Anangu Pitiantiatiara (A P) Lands

The Anangu Pitjantjatjara lands compdse 10,190,000 hectares of semi-arid land or

approximately 1Oo/o of South Australia (see figure 3.2.1.f ). The lands are part of a

central plateau that extends from the far north-west of South Austral¡a into the

Northern Territory and across into Westem Austral¡a. The plateau is approximately

700 metres above sea level. The Anangu Pitjantiatjara lands are dissected by four

principal ranges: the Musgrave, Mann, Tomskin and Everard, all of whiCh run East'

West and dse to 1,300 metres, with some peaks over 1,600 metres. Mt Woodroofe in

the Musgrave Range is the highest peak in South Australia. The area receives, on

average, 20-25cmof rain per year, so the watercourses are generally dry, and untilthe

provision of bores water was only available lrom rock holes and soaks found in the

ranges. The summer is usualty hot with temperatures averaging 35+oC, winters are

typically cool by day with cold nights.

ln 1976 the P¡tjantjatjara, Yankunyatiara and Ngaanyatiara people formed the

P¡tjantjatjara Council ¡ncorporated to represent their c-ollective ¡nterests. lt operated

98 Chapter 3 across state borders, administered some human and essential services and oversaw land managemenl systems. ln 1981 the South Australian Government, under the

Pitjantjatjara Lands Rights Act, gave freehold title of the land to the Anangu

Pitjantjatjara (Nganampa Health Council 1 987).

Yalata Lands

Lying some 1,500 kilometres lo the south of the Anangu Pitjantiatjara lands lie the yatata lands. The community of Yalata lies a few kilometres otf the Eyre Highway and approximately 170 km west of Ceduna, on the edge of the Nullabour Plain. The plain has an even and treeless surlace, is noted for its deep flooded sinkholes and caves, and for high cliffs and sand dunes on its coast.

The other half of the Yalata lands extend north to Oldea and further to the Maralinga lands (taking in some of the Great Victoria Desert). These areas contain open mallee,

some grassland and scrubland typicalof the semi-arid zones which receive less than

2OO-300 mm of rain a year. Like the A P Lands the summers are hot and dry with the

temperature during the winter being moderate in the dayt¡me but cold at nighl. There

are oceanic influences in the southern areas such as coaslal fogs (Atlas of South

Australia).

oo Chapter 3

Figure 3.2.1.1

The Anangu Pitiantiatiara and Yalata Lands 1990

o Alice Springs

NORTHERN TERRITORY

Wingel tnna J (Pukatja) I

(lwantja) !

I

I

I o Coober Pedy I SOUTH ! I AUSTRALIA I

i YALATA I

Ceduna i

i

I

i

I l'' -'

o Adelaide I

I

i

I

I

I

I

100 9.2.2. Descrlptlon of the Study Populatlon

The population of the Anangu Pitjantjatjara and Yalata lands is predominantly Aboriginal.

However, a significant population of white people provide services in health, education, maintenance and management. Although some are permanent residents in the communities, most spend only 1 - 3 years on the lands.

major ln the Anangu pitjantiatjara lands the majority of the Aboriginal population live in the commun¡ties; however, a proport¡on live in "homeland" settlements scattered around the who major communities. These homelands otten have specific associations with the families the occupy them, and have been established over the past 10-15 years at the instigation of

Anangu Pitjantjatiara.

The population of the communit¡es and the lands as a whole is dynamic. Movement of people

between the communities and homelands, and further away (such as west to Warburton,

north to Alice Springs, or south to Port Augusta) occurs frequently. Although families ¡dent¡fy

strongly with a given area, individuals may spend months or years ¡n other commun¡t¡es. This

mobility can interfere with the accuracy of single census. Table 3.2.2.1 details the population

distribution in the two areas according to the 1986 census (Australian Bureau of Statistics

t9g6 Census), an estimate based on population lists compiled by Nganampa Health Service

(section 4.3), the population distribution of the sampte from this study, and a projection of the

1g86 census data lrom 1986 to 1990 (Fig. 3.2.2.1). The proiected distribution assumed no

net migration, an even distribution of individuals within an age-group, used the populalion list

estimate for the 0-4 age-group and used an estimate for the mortality rate ¡n the 50+ age-

group from the Australian Bureau of Statist¡cs (appendix 2). The chief purpose of the

projection was to indicate the difference between the population estimate as derived by

Nganampa Health Service and that obtained by the census. Either there had been

considerable net migration ¡nto the lands, or the 1986 census underestimated the population

tolal, or the poputation list est¡mate was exaggerated. Given the difficulties, both with the

single census estimate due to the mobility of the population, and the methodolog¡cal

101 Chapter 3 problems assoc¡ated with the population estimate lrom the Nganampa Health Service, the population totalfor 1990 probably lies somewhere between the two estimates. However' a with survey in Ernabella found that only 10% of the population list was no longer associated that community (section 3.3.2.). Therefore, the population list compiled by Nganampa Health service, is probably the most accurate estimate of the true total for the source population and was used ¡n data calculations'

Table 3.2.2.1: Population Distribulion of the A P and Yalata lands.

1 986 1986 Census Population 1989/90 Census data proiected List 1990 Survey to 1990

Age-group MFMFMFMF

0-4 115 106 165 162 165 162 99 95

5-9 88 98 110 126 195 220 121 152 174 1 0-19 208 2',12 196 206 297 313 125

20-29 184 172 194 188 256 267 86 135

30-39 89 94 127 126 135 180 55 92

40-49 7A 60 82 74 121 110 65 67

50-59 56 52 51 49 96 92 52 60

60+ T3 67 71 76 8rÍl 97 62 74

891 861 996 1006 1348 1441 665 849

total 1752 2002 2789 1514

102 Chapter 3

Figure 9.2.2.1.: Population distribution lor the A P lands

Females 1986 Census ø ¡ Males 400 1990 PoPulation a Females Males List I

300

Number of lndividuals 200

100

0 04 s9 1019 bæ 0æ 40,49 5050 æ+

Age-group

103 Ghapter 3

3.2.3. Descrlptlon of the Collaboratlve Organlsatlons

The South Australian Aborigi nal Trachoma P rogram

was established in The South Austratian AboriginalTrachoma Program (Trachoma Program) now has a full- 19g1 as the S.A. Eye Care Program. lt was formally constituted in 1983 and five members and a time coordinator and administrator, with a voluntary executive body of the chaiçerson. Once ayearthere is a general meeting between it and the representatives of

Aboriginal commun¡ties it services.

South The program provides ophthalmic care services to Aboriginalcommunit¡es throughout They also Australia, usually by annual visits of ophthalmologists or eye health educators. other eye conduct school screenings and individual examinations to detect trachoma and

disease, and conduct training programs for Aboriginal health care workers'

Nganampa Health Seruice

people the The Nganampa Health Service is controlled by the Anangu Pitjantjatiara through the A P Nganampa Health Council (N.H.C.). lt provides health services to the communities of runs medical lands in the north-west of South Australia. Atthough based in Alice Springs it

with a mifure and dental services in the communities. lt provides a culturally sensitive service The of Aboriginal health care workers and usually non-Aboriginal medical and nursing staff.

performed A P lands' N.H.C. also determines what research related to health can be on the

Atu rigi n al H e alth O rg ani satio n

The Aboriginal Health organisation (A.H.O.) is an incorporated body under the south

Australian Health Commission Act ol 1981.

It has a principal objective of increasing the health status of Aboriginal people in South of Australia, and has specific policy goals which include: understanding the health needs

Aboriginal people in s.A.; the creation and maintenance of health care services to the

104 Chapter 3

provision Aboriginal community, and the involvement of Aboriginal people in the of these health services

The A.H.O. has adopted several initiatives to achieve its objectives these include: the Health establishment of community-controlled health services; the training of Aboriginal data workers, the provision of hospital liaison officers; the establishment of a comprehensive and advice to base of all aspects ol Aboriginal health in S'A'; and the provision ol ass¡stance Aboriginal departments organisations proposing to carry out medical research in the community. A formalty constituted Ethics committee was established in 1991.

The A.H.O. is not lormally connected with either the Nganampa Health service or the

Trachoma program but does liaise with both organisations. The Trachoma Program meets regularly with representat¡ves of the N.H.C. and coordinates through the Nganampa Health service, visits of ophthalmologists to the Anangu Pitiantjatiara lands.

105 Chapter 3

3.2.4. StudY Deslgn

The study was divided into lour phases

Planning and Facilitation

2 Fieló¡¿ork

3 Collation and Data Processing

4 lnformation Distribution

well ln a study of this kind the choice of survey type was limited. Ophthalm¡c disorders are

suited to epidemiological studies (Rosenthal 1988), and a descriptive cross-sect¡onal survey

was chosen as the most appropriate form. The most comprehensive picture would be

obtained from a census-type survey, in which every member ol a community is enumerated

and his/her ocular status recorded. A less comprehensive, but still standard method, uses a

simple random sample which assumes that there is a way of identifying individuals within a

community e.g. elecloral rolls or school l¡sts. The lists should be as comprehensive and up to

date as possible. A sample size is calculated on the basis of the est¡mated prevalence for the

condition under consideration, and people are picked at random from the list to make up the

sample. A third way to conduct a survey is to use a random clusler technique, which assumes

that there are distinct groups ol people with homogeneity in the factor of ¡nterest between'

and heterogenity within, clusters. Groups are picked at random from a total populat¡on

representing, for example, communities, schools or other easily identified groups. This

smaler sub-set of the population is then surveyed and the resulls extrapolated to the total

population. lt is assumed that the clusters are represerìtat¡vs of the population from which

they ars drawn.

It was apparent in this study that there were not enough distinct communities that were

sufficien¡y comparable to rely on cluster sampling. A simple random sample could not be

performed because population lists were not known to exist. The census style of survey was

chosen because it was theoretically achievable and previous surveys had used the same

technique. The results would thus be comparable, and it also enhanced the service role of

r06 Chapter 3

of the research in case identilication and treatment. The service role and educational aspects the research project were key elements ¡n the survey's eventual approval by the Aboriginal communities previous research had sometimes led to misunderstandings and illfeeling when work of little direct benefit to the people was pelormed. communities, or organizalions representing The them, began to evatuate any research proiect for its immediate and long term benefit.

Anangu pitjantjatjara had been subjected to numerous surveys and research in the 1970s in and early 19g0s. A renal survey had been conduc{ed in the north-west of South Australia

19g2/g3 by The Aboriginal Heatth Organization of South Australia (A.H.O. Renal Survey

lggg). lt was culturally sensitive, provided a significant service role and gave cont¡nu¡ng

feedback to both the individual participants and the communities. The design and conduct of

the Renal Survey provided a basis for the fornrulation of this eye heatth survey.

The research proposal was odginally written as an exercise for course work in a Masters of

public Health Degree (University ol Adelaide), and was produced in the lormat of a National

Heatth and Medical Research Grant (N.H.&M.R.C.) grant proposal. lt was subsequently

reworked after comments from the markers to make it more relevant to the ethical and health

councilcommittees. lt included a review of the relevant trachoma and blindness l¡terature, as

well as a study design and research plan.

Funding was not sought for the project as it was being pelormed under the auspices of the

S.A. Aboriginal Trachoma Program, which had agreed to províde travel expenses, as well as

logistical and specialist back-up. The survey was conducted as an extension of the normal

trachoma program work.

As part ol the initial work, and prior to the survey, the principal investigator spent several

sessions in fhe eye outpatients' department of Flinders Medical Centre, refreshing basic

ophthatmology skiils and knowledge. The new lrachoma grading system (5 sign system) was

reviewed utiiizing photographs and slides (W.H.O. 1989). A pilot survey was conducted at

point pearce, an Aboriginat community on Yorke Peninsula, South Australia, in Apñl 1989.

107 Chapter 3

procedures were Examination lechniques, trachoma grading, data collection and recording evaluated. These data were not included in the subsequent survey'

3.2.5. Ethlcal APProval

Human Ethical approval was sought from four bodies. The Committee on the Ethics of Flinders Experimentation (University ol Adelaide) and the Clinical lnvestigatíon Committee at An Medical Centre (the place of employment of the ¡nvestigator), were approached. advisory ethics committee of the Aboriginal Health organisation, although not formally organized as an and committee at that time, provided a vetting and advisory role which was invaluable. Finally most importan¡y, Nganampa Health Committee, a body that acted lor Nganampa Health

Councit and the Anangu Pitjantjatjara was approached for approval. Only with the Nganampa

Health Commiüee's approval could the survey be undertaken, and permits to enter the

Anangu P¡tjantjatjara lands obtained'

Separate approval to visit and pelorm the survey on the Yalata lands was obtained from the

Yalata CommunitY Council.

Approvalwas given by each organisation with minimal alterations to th€ protocol (appendix 5).

3.2.6. The SurveY

Description

The source population was the Aboriginal population of the Anangu Pitjantjatiara and Yalata

lands in South Australia.

lnitial target population estimates were based on 1986 Census data. On reaching the

commun¡t¡es it was lound that more recent population lists had been prepared and then

updated in lggg/1gg0 by Nganampa Heatth and Yatata medicalservice statf. Medicalrecords,

108 Chapter 3

and ths pension and social security recipient records, school lists, community wod< records provide l¡sts of knowledge of local Aboriginal health workers were combined to alphabetical the people in each community ( Table 3.2'6.1).

of the A cross-seclional survey was undertaken. Using the population lists all members to community were approached to part¡cipate in census style. Each rrerson who agreed participate was examined and had their visual acuity tested. Those with specilic visual problems were referred to an ophthalmologist (see The Examination).

Population Lists

As previously described, this sludy coutd not check the accuracy of the population lists; however, the lists were useful adjuncts to the study. They were reananged into sex and age- specific categories and used:

1 . as a check-list lor individuals, names and ages;

2. as a guide to the population distribution;

3. to ¡dentity individuals not previously seen in a community

people who ln each community the tists were updated by Aboriginal health workers to identify

had moved e¡ther into or away from the community. The names of newly bom children were list. added to the list and, although infrequent, those who had died were removed lrom the

109 Chapter 3

Table 3.2.6.1: Total population in the communities:l9$1 census, 86 census and 1989/90 PoPulation list data

1989/90 Community 1 981 1 986 Census Census Population Data List

586 Emabella 322 364 Fregon 203 268 307 lndulkana 301 237 360 Mimili 132 144 215 437 Amata 180 276 1t Kalka/Pipalyatiara eg 105 330 Yahta 281 222 ) 2 Oak Valley 118 ) 554 3 Half Way Canp 15 18 )

Total 1557 1752 2789

1 Estimate from available 1976/1981/1986 data.

2 No census figures for 1981 Oak Valley ( not populated then).

3 Estimate based on 1989/90 figures.

Sample Size

Although the survey was to be of a census type, a sample size calculation was required to

confirm that the source population was large enough to g¡ve adequate precision to an guide estimate of the prevalence of trachoma and visualimpairment. lt would also acl as a to

the etfort, time and cost involved. A sample size was catculated using the following formula

(Schaeffer et al. 1990).

110 Chapter 3

n 4.N.(P).(1 - P)

(N - 1) 82 + a.(P).(l - P) where n = sample s¡ze ¡ = population size B = 2standard error bound on the estimate of P p = population proportion of interest

The population size was estimated using 1986 Census data projec{ed to 1990, a figure of population 2002 was derived (Table 1 .1). The error bound of 0.5% was selected and the .Red proportion of ¡nterest was the rate of binocular blindness in the Centre" zone of the

N.T.E.H.P. survey, which was 2.3% for all ages (R.A.C.O' Report 1980). Substiluting

fl= 4.2002. (0.023).(0.977)

2 (2002).(0.005) + a.(0.023).(0.e77)

n 179.9

0.1399

1 286

A sample size of 1,286 individuals would be required to est¡mate the proportion of ¡nd¡v¡duals

binoanlarly blind in the population with an error o10.5%.

A similar calculation for trachoma was performed using N = 2002, B = 5.07o, P = À57o (taken

lrom the R.A.C.O. Report 1980 for c¡catricialtrachoma). A sample size of 4Í19 individuals was

required to estimate the proportion of individuals with lrachoma with an error ol 5.070'

111 Chapter 3

Transportation and Dates of Fîeld Trips

service Dates for individualfield trips were made after discussion with each community health lands and the S.A. AboriginalTrachoma Program. Entry permits for the Anangu Pitjantiatiara were then obtained. ln the case of Yalata, permission was obtained directly from the community council.

I chartered aircraft or the Royal Flying Doctor service (R. Augusta) were used to transPort

personnel and equipmenl to the commun¡t¡es.

The survey team comprised the survey researcher, an ophthalmologist, and one or two

helpers (e.g. Aboriginal nurse educator, trachoma program coordinator, medical student, or

additional ophthalmologists).

A totalof eight lield trips were undertaken.

Date Community Duralion

1989 July Ernabella 5 days 1989 August Pipalyatjara/Kaka 5 days 1989 September lndukana and Mimili 1 2 days 1989 October Yalda 9 days

1990 May Emabella and Amata 18 days 1990 June Fregon and Kalka 14 days 1990 August Yalata 5 days

These were the major communities for the Anangu Pitiantjatiara in these areas. Smaller

satelt¡te homelands around these population centres were also visited, and included ¡n the

data collected for these communities.

visiting rernote areas. 1 . Using a S.A. Health Commission funded aircraft lor specialists

112 Chapter 3

The Examination

posters prior to the arrival of the survey team, the community's health service advertised by and word of mouth that the Trachoma Program team would be visiting. All the commun¡ty members were encouraged to attend.

position in room with A suitable area was found at each clinic, either outside in a sheltered or a slit adequate illumination. Another area was found to provide an area suitable for ophthalmic, lamp or fundoscopic examination ol the eye. on their arrival participants would be greeted and introduced to the researcher. The survey E¡thor and testing procedures would be explained. Verbal consent would then be sought. the part¡cipant or researcher would write the part¡c¡pant's name in the top lett hand corner of the data sheet, their age and sex were then entered. lf e¡ther the age or name were unclear, the population lists were consulted, with either the Aboriginal heatth worker or the individual

participant identifyi ng the appropriate ¡ nf ormat¡on.

An illiterate Snellen E chart was used for visual acuity testing, with the subiect at six metres.

covering first the left eye then the right eye, the subiec{ indicated by a hand movement lhe

direction in which the legs of the E po¡nted. The tester noted any signs of strabismus. A

conected vision with glasses or pin hole was also obtained il required.

Using x2 loupes the eyes were examined for basic eye problems using a bdght torch or

ophthalmoscope. External signs of trichiasis or comeal opacity were sought. The lids were

then everted, and examination of the upper tarsal plate perlormed to look for active trachoma,

or trachoma scarring according lo the five sign system (see Appendix 4)'

Referralto the ophthalmologist occurred if

6/18 in any eye; 1 . repeat test¡ng of corrected visual acuity was worse than

2. refraction was required;

113 ChaPter 3

3. signs of trichiasis or corneal opacity were evident;

4. there was any abnormality ol the eye which was thought to require a specialist

oPinion;

5. the participant had any chronic disease which might have ocular

consequences (e.g. diabetes or hypertension);

6. the participant requested a specialist opinion'

pupils, used a The ophthalmologist repeated the previous examination, dilated both and portable slit lamp and direct or indirect ophthalmoscopy to make a diagnosis, based on one of several categories (APPendix 3).

3.2.7. Data Handllng

Recoding Data

pre-printed lorms At the time of the examination data relating to an individualwere entered on

(Appendix 3). The forms were adapted from a concurrent eye survey ol blindness in South

Australia.2 Some questions were redundant and the information not sought. Names were

added for the Aboriginal population to assist in cross-checking part¡cipants between for communities, and to ensure that the sheet could become part of an ocular medical record

that individualwhen it was returned to the community medical service.

Data were recorded in the following sequence: name, age, sex, past ocular h¡story, vision,

results basic eye examination, previous eye surgery,lrachoma, and then optional examination were also if seen by the ophthalnologist. causes of low vision and the current action required

noted.

2. Conducted bY Dr. H. Newland.

114 ChaPter 3

ticked, names At the end of each day, all forms were checked to ensure all boxes had been had to be entered clearly, and age and sex recorded. At t¡mes it was found that ages totals were corrected after examination of the population lists or medical records. Running

kept in age-specific groups for trachoma and causes of blindness.

all individuals At the end of a commun¡ty visit, the medical service was provided with a l¡st of

requiring:

1 . treatment for trachoma;

2. surgery for cataract or other conditions;

3. follow up for certain conditions, e.g. trichiasis, lrachoma scarring, diabetic

retinoPathY.

The community councilwas also given non-identifying clinicalinformation and tabulated data;

including the total number seen, grouped by age and sex, and the number requiring trachoma

lreatment, surgery or the prescription of spectacles'

Data sheets were given community and individual identity numbers and entered onto a dala-

base using Epi-lnlo (5), a sottware package lor the analysis of epidemiological survey data.

Data Check¡ng

There were severa¡ stages of data checking:

1. Field checks were undertaken during the survey'

2. Dala sheets were arranged alphabetically for each community and checked

for duPlicates.

g. printouts ol all inlormation entered into the computer were obtained and

compared with individual data sheets after computer data entry, and appropriate

corrections were made.

115 Chapter 3

4. Cross tabulations and frequency tables were used to identify outlying data was points and missing data. Comparison with field (manually calculated) data also

performed.

S. lf missing age or sex data could not be obtained, the information lor that or individual was not entered. Missing values for visual acuily, eye abnormality

trachoma status were left as missing but any other information was entered.

retained for When duplicate examinations were compared the most recent data sheet was error (section data entry, and the l¡rst sheet used to assess inter-observer and ¡ntra-observer s.2).

Data Analys¡s

frequerrcy Analysis was pelormed using Epi-info 5.0 (a statistical package), including simple allow tabulations for descriptive assessment of the results. Age/sex standard¡sation to comparison between communities and between ditferent surveys. Odds ratio calculations using Mantel-Haenszel X2lo¡ trend to quantitate and test differences between serial surveys' when the age-groups and the sexes. The Fisher exact test was used to est¡mate the P value expected cell frequency in one or more cells of a 2x2 table was below 5.

Logistic regression analysis was performed using Systat (a stat¡st¡cal package) to investigate

the etfect of several variables on the prevalence of trachoma'

Community Liaison

ln severalc¡mmunities the community council would be consulted on the first or second day

of of the visit. The survey team would be introduced and the purposes and expected benefits

the survey explained. Advice would be taken from the council concerning visits to houses

within the community, and visits to outlying camps or homelands. This was ol particular

relevance when there had been a recent death in the community and people m¡ght be ln

mourning, or if there were special sites that could not be visited. Also, the occurrence of

116 Chapter 3

minimize religious ceremonies in the latter months of the year were able to be discussed, to aims ol the any interference by the researchers and facilitate the accomplishment of the survey

9.2.8. lnformatlon dlstrlbutlon lnlormalion distribution occurred at three levels. At the end ol each visit, the health service council was given clinical information pertaining to individual participants, and the community was given non-identifying tabulated data. On the investigato/s return to Adelaide the South

Austratian Aboriginal Trachoma Program (Trachoma Program) was given copies of all the aggregated data for each community. Summary reports were prepared at intervals during survey period.

given When the survey work was completed, an interim report was produced and to passed Nganampa Heatth Council (N.H.C.). Atter consideration by N.H.C. the report was to the Trachoma program and the S.A. Aboriginal Health Organisation (A.H.O'). A bdef summary

of the survey work was g¡ven to tho eth¡cal committees of the University of Adelaide and

Flinders University of South Australia.

given N.H.C', A final report, detailing all resutts and comparisons with previous surveys, was to

the Trachoma Program, and the A.H.O.

Any future articles wfitten for publication will be forwarded to N.H.C. for comment and

approval.

117 Chapter 3

3.3 Results

3.3.1. DescrlPtlon of the SamPle

population; approximately A total of 1514 individuals were examined out of the entire source four times) either 200 were seen twice (and in the case of one school age boy, inadvertently, a ditferent on the same or a subsequent visit to the community, or at another community on

occasion. The duplication was established either at the t¡me of the examination or of the subsequenily at the time of the data sheet checking. The age and sex distñbution

sample (the study population)obtained is outlined in Table 3.3.1.1 and fig'3'3'1'1

Table 3.3.1.1 : Age and sex distribution for the 1989^90 survey sample

Age-Group Females Males Total %Total

0-1 29 26 55 3.6 2-4 oo 73 139 9.2 5-9 152 121 273 18.0 12.0 1 0-14 100 82 182 I 5-19 74 43 '|.17 7.7 20-29 135 86 221 14.6 30-39 92 55 147 9.7 40-49 67 65 132 8.7 50-59 60 52 112 7.4 60+ 74 62 136 9.0

Total 849 66s 1514 100 % of total 56.1% 43.9o/o

The table shows that 42.8% of the sampled population was under 15 years of age, and that

females formed 56.1% of the total sample. The largest part¡c¡pat¡on rate difference between

males and females occurred in the 15-19 and 30-39 age brackets. The distdbution should be

118 Chapter 3

in which females comprised compared to rhe underlying source population (section 3'2.2'1,

51.7V" of the total.

Figure 3.3.1.1: The population distribution of the sample

200

ø Females I Males

Number of

100 lndividuals

0 01 Z4 5.9 1û14 l$19 bæ Sæ 40'-4S 5059 m+

Age-group

119 Chapter 3

3.3.2. PartlclPatlon Rates

visit' participation rates ¡n each community varied due to several factors: the duration of the importance the the occurrence of ceremonies, business or sporting events and the community and individuals placed on the survey (section 5.2.1.1.

population in each community' Table 3.3.2.1. ouilines the percentage of lhe source surveyed found in any It was never expected that all the names on the population list would be of people who community at a given time. Rather the population list reflecrted the number year the shopping and identified with that community, stayed there for much of the or utilized

other facilities on a regular basis (e.g. people from the surrounding homelands)'

validity of the A small population survey was conducted in Ernabella in 1990 to check the

population list. Utilising the knowledge of the local health workers, the ¡nvestigators examined

the population lists. lndividuals who had not been seen were idenlified, and their moved whereabouts ascertained. particular care was taken to ensuro only those who had of the population list away for a "long time, were counted. lt identified 61 individuals, or 10.47" Port as having moved away permanently or semi-permanently, usually to Alice Springs'

Augusla or to other commun¡ties within the Anangu Pitjantjatiara lands'

who were During the survey a small number of individuals (approximately 20) were examined otlered visiting lrom adjacent communities or towns. Although checked for eye disease and group identified, treatment, their results were not included in the survey. There was another population lists. They those who lived permanently on the lands and did not appear on the population list totals were added to the l¡sts when found and therefore were included in the individuals were used in this and other chapters. No attempt was made to quantity how many been bom in added ro the lisls; however the majority were babies and young ¡nfants who had people the previous year. There were only 15 adults who had to be added to the l¡sts, usually

who had recently moved onto the lands.

120 Chapter 3

Table 3.3.2.1.: Community participation rates

o/o 1990 PoPulation 1989/90 Lists Survev

151 Yahta F 274 M 280 144

Total 554 295 s3.3

107 Fregon F 153 M 154 87 30 I 4 6

Emabella F 321 224 M 265 171 58 95 67. lndulkana F 182 108 M 178 74

Totel 360 182 50.5

Amata F 223 103 M 214 81

Tota¡ 437 r 84 42.1

F 116 60 Mimili M 99 33 Total 215 93 4 3.3

Kalka/ F 172 96 Pipalyatjara (PiP) M 158 75

Totel 330 171 51.8

AllCommunities F 1441 849 M 1 348 665

Total 2789 1514 s4.3

time was The poorest response occurred ¡n the two communities in which the least amount of in Emabella where a spent, Amata (4 days) and Mimili (3 days). The best response occurred

total of 19 days were sPent.

121 Chapter 3

was approximately the The male to female ralio of the source population across communities Mimili showed same, the exception was Fregon, where there were more males than females'

the targest female to mate participation rate difference ol65% vs 35% respectively.

demonstrates Of interest was the variation in participation across age-groups Table 3.3'2.2

this variation.

Table 3.3.2.2 (A): Participation rates across age€roups Females

Age-Group Yalata Fregon Ernabella lndulkana Amata Mimil Kalka/Pip Total

0-4 15t23 11/16 24t37 15t28 11t26 11t17 8/15 95/162 65% 69"/" 65% 54o/o 42o/" 65"h 53"/" 59 % 1521220 5-9 28t32 11t17 50/55 18t32 17t37 10/18 18t29 88% 65% 91% 56% 46"/0 56% 62% 69%

10-19 40t76 24t38 47t74 24t31 17t42 8117 14t35 1741313 53% 63% 64"/o 77% 40o/" 470h 40/" 56"/o

20-29 21t46 17t26 41t62 18/31 17t45 12t27 9/30 1351267 46% 65% 66% 58% 38o/" 44o/o 30"/" 51%

30-39 18/46 14t20 22t35 8/16 13t32 4111 13t20 92/r 80 39% 70o/o 63% 5O"/" 41Yo 360/" 65% 51o/o 10 40-49 1 0/18 8112 14t23 7t14 6/10 6l'|.2 16t21 67/f 56% 67% 61o/o 50% 60% 50"/" 76o/" 61 %

50-59 10t14 8/1 0 17t23 8117 8114 3/6 6/8 60/92 71% 80% 74% 47% 57"/o 50% 75% 64%

60+ 9/19 14114 9112 10/13 14t17 6/8 12114 7 4197 47% 100% 75% 77Io 82o/o 75% 86% 7 6.3%

Participation by females varied lrom 100% in Fregon in the 60+ age€roup, to a low of 30% ¡n

2O-2g age-group in Kalka/Pipalyat¡ara. The best part¡c¡pat¡on occurred in the 60+ age-group

which had the most eye prob¡ems and therefore was ¡nterested in participating in the survey.

The school age-group 5-g, also showed above average part¡c¡pat¡on, probably because they

were easily found at school. The ditferences between the part¡cipat¡on of older age€roups

and the middte years probably reflected the mobility of the younger people and their relatively

good eye health.

122 Chapter 3

Table 3.3.2.2 (B): Participation rates across age{roups Males

Age-Group Yalata Fregon Ernabella lndulkana Amata MimiE Kalka/Pip Total

99/165 0-4 17t27 12t20 19t25 12t26 17t35 9/f 6 13/16 60"/o 63o/o 60% 76% 46% 49% 56% 81"/" 121/195 5-9 24t36 17t25 37t51 9/16 12t26 7t17 15t24 67% 68% 73% 56% 46% 41"/" 63% 62"/o 1251297 10-19 32t68 9t29 44t58 15t41 8/49 3117 14/35 42o/o 47o/" 31% 76% 37"/" 16o/o 18o/o 40o/o

20-29 23l56 11t23 28t55 8134 8/38 2t20 6/30 86/256 41% 48o/o 5',1v" 24o/" 21Yo 10% 20"/" 34o/o

30-39 12t33 7t11 8122 't0t20 10t23 5/1 0 3/16 55/135 41o.h 36% 64o/o 36% 50/" 43"/" 50% 19%

40-49 18t24 13t20 7t18 7t20 5/1 3 2t8 13/18 651121 75% 65/" 39% 35% 38"/" 25% 72"/" 46o/o

50-59 11t23 7t13 17t23 4t10 7t12 3/6 3/9 52/96 48% 54% 74o/o 40ø/o 58% 50% 33% 54"/o

60+ 7t13 11/13 11/13 9/1 1 14t18 2t5 8/10 62183 54% 85/" 85% 82"/" 78% 40% 8OVo 7 5o/o

(60+ age- For males the participation rates varied from a high of 85% in Fregon and Ernabella

group),toalowof 1O%inMimili (20-29 age-group). Overall thepatternparalleledthatfor The females, w¡th greatest participation occurring ¡n the 60+ and 5-9 years age-groups'

occurred ¡n differences between fhe 60+ and 20-39 age-groups was greater than that which

females, probably due to the greater mobil¡ty of the young men compared lo young won¡en'

123 Chapter 3

3.3.3. Overvlew

years of age. The illiterate Visual acuity was tested in all individuals except children less than 6 movements to indicate the snellen E chart was used al6 metres, with the subject using hand direction of the E. Visual acuity better than 6/6 was not tested for.

which was recorded for all A pinhole or glasses were used to obtain the correc{ed visual acuity, individuals with less than 6/6 vision in e¡ther eye visual acuity was divided into ten categories lor coding pufposes.

0 6/6 uncorrected

1 6/6 corrected

2 6112 corrected 3 6/1 8 corrected 4 6/36 corrected

5 6/60 corrected 6 3/60 corrected 7 CF conected (Count ligures at I metre) I PL corrected (PercePtion of light) o NPL corrected (No perception ol light)

The tables Visual acu¡ties for males and lemales are presented in Tables 3.3.3.1 and 3.3.3.2. the age of 6 represent the best vision obtained by each individual participant. Children under was abnormal. were assumed to have 6,t6 vision in both eyes unless the oq¡lar examination

124 Chapter 3

Table 3.3.3.1: Right vs left eye corrected visual acui$ for males

LCORRECTED

6/6 6/6 6112 6/18 6/36 6/60 3/60 CF PL NPL Total Uncorr. Corr

RCORRECTED

0 559 6/6 Uncorr. 559 0 0 0 0 0 0 0 0 1 31 6/6 Con. 0 15 6 2 3 1 0 2 I 26 6t12 0 5 14 2 2 1 0 0 2 0

1 2 16 6/1 I 0 2 2 4 2 3 0 0 1 1 11 6/36 0 1 1 2 1 3 0 1 1 5 6/60 0 0 0 1 2 1 0 0 0 3/60 0 0 0 0 0 0 0 0 0 0 0 CF 0 5 0 0 0 0 0 3 0 0 I

1 5 PL 0 0 1 1 1 0 0 1 0 0 0 0 4 NPL 1 1 2 0 0 0 0

Total 560 29 26 12 11 9 0 I 4 6 665

of 665 Corrected visual acuity ol6112 or better in both eyes was achieved by a total of 599 count males (90%). Only live or 0.75/" of males were binocularly blind. (W.H.O. definition:

fingers at 1 metre or worse). Broadening the definition to less than 6/60 (Australian blindness

definítion) did not add any more individuals to the total.

125 Chapter 3

Table 3.3.3.2: Right vs lett eye corected visual acuity for females

LCORRECTED

6/6 6/6 6t12 6/18 6/36 6/60 3/60 cF PL NPL Total Uncon. Con.

RCORRECTED

0 711 6/6 Uncorr 711 0 00 0 0 0 0 0 2 0 2 1 2 30 6/6 Corr. 1 18 2 2 0 1 I I 24 6112 0 5 I 5 1 2 0

1 2 1 24 6/1 I 0 5 I 6 0 I 0 14 6/36 0 2 0 2 2 1 0 4 3 0

1 1 0 0 9 6/60 0 0 1 1 4 1 1 3 3/60 0 0 0 1 0 0 0 1 0 2 19 CF 0 5 3 1 3 0 0 5 0 9 PL 0 0 0 2 1 1 0 2 2 I 1 6 NPL 0 0 1 0 2 0 0 1 1

Total 712 35 29 20 13 I 1 18 10 9 849

(W.H.O. definition)' By comparison with the male results 15 women were binocularly blind or better representing 1.77%of the total (cf 0.75o/o in males). Corrected visual acuity ol6112

was found in 745 individuals or 87.7o/o (cf 90% in males). These diflerences were not The prevalence explained by vadations in age structure and must relate to some other factor. men' and of severe trachoma, eg. trichiasis and comeal opacity was greater in women than prevalence probably accounted for much of this dilference (Tables 3.3.7.4 and 3'3'7'5)' The

greater women than ¡n of sight limiting cataract (Tables g.g.4.2 and 3.3.5.2) which again was in

men, would also be a cause of this difference.

126 ChaPter 3

3.3.4. Blnocular Bllndness

with Australian Two definitions of binocular blindness were used;the first to allow comparison

dala, and the second to meet W'H.O. cdteria.

less than 6/60 Australian definition - those individuals whose vision in their better eye was

(Social Security Department 1 991 ).

than count W.H.O. definition - those individuals whose vision in their bener eye was no bstter

fingers (C.F.) at one metrs (W.H.O. 1979).

Table 3.3.4.1 gives the number of individuals binocularly blind by age'grcup and sex.

127 Chapter 3

Table 3.3.4.1: Binocular blindness sex by age-group

Australlan defn.:

o/o Total % Age- Females "/o Males Group

0.0 0-19 01421 0.0 0/345 0.0 0/7 66 0.9 20-29 2t135 1.5 0/86 0.0 21221 0.0 30-49 0/1 59 0.0 0t120 0.0 o1279 0.9 50-59 1/60 1.7 0152 0.0 1 1112 I 4.0 60+ 14t74 18.9 5t62 8.1 19/136

Toral 171849 2.O 5/665 0.8 2211514 1.5

W.H.O. defn.:

o/o o/o Age- Females % Males Total Group

0-19 0t421 0.0 0t345 0.0 0/7 66 0.0 20-29 2t135 1.5 0/86 0.0 21221 0.9 30-49 0/1 59 0.0 0t120 0.0 ot279 0.0 50-59 1/60 1.7 0t52 0.0 11112 0.9 r 60+ 12t74 16.2 5t62 0.1 17l136 2.5

Toral 15/849 1.8 5/665 0.8 2011514 1.3

using lhe A total of 20 individuals, or 1.3"h of the sample population were binocularly blind

W.H.O. delinition of blindness

The est¡mate of the age-standardized odds ratio (Mantel-Haenzelweighted OR) of blindness' oR=3.22,95o/o using the Australian definition, when females were compared to males was:

C.l.: 1.03 to 10.4Í! , (X2=4.04 (p=0.0a$).

128 Chapter 3

(a) Causes of Binoaiar Blindness:

have been The causes of binocular blindness were grouped. Only the maior categories maior presented; when two or more causes have been given lor binocular blindness the contributing cause has been recorded (Table 3'3'4'2 )'

(45%). Cataracl Binocular blindness was caused by trachoma in nine ol twenty individuals and congenital accounted for another eight blind individuals (4}o/ol, trauma, infection problems the remaining 15%.

was not severe Diabetic retinopalhy, atthough causing visualimpairment in some individuals,

enough yet to classify these individuals as blind by w.H.o. standards.

'129 Chapter 3

Table 3.3.4.2: causes of binocular blindness (w.H.o delinition)

Female

Cause

Age- Absent or Catarad Trachoma Other Total Blind Group phthisical globe

0 1 0-14 11 2 20-29 1 30-39 0 40-49 0 50-59 1 1 12 60+ 1 4 7

Total 2 4 I 1 15

pupilwith cataract' 1 Right total retinal detachment and left congenital ectopic

Male

Cause

Age- Absent or Cataract Trachoma Other Total Blind Group phthisical globe

20-29 0 30-39 0 40-49 0 50-59 0 60+ 4 1 5

Total 0 4 I 0 5

130 Chapter 3

that trachoma and cross-tabulation of the cause of blindness for left and right eyes reatfirms (Table The causes cataracf were the leading causes of blindness for individual eyes 3'3'4'3)' is also presented for of blindness for those with visual acuity ol 6/60 or less in both eyes comparison with the 1976-1979 N.T.E.H.P' survey'

Table 3.3.4.3: Causes of blindness right vs lett eye for visual acuity of

CF at 1 metre orworse (6/60 orworse in brackets)'

Lett Eye

Absent or Cataract Trachoma Other phthisical

Absent or 1 phthisical I 3 Right Cataract 7(21 (1) 1 Eye Trachoma 3(1) 1(1) s(2)

2 1412¡5,6 Other 1

1. One case of presumed radiation induced bilateralcataract

2. Right cornealoPacitY

3. Left cornealoPacitY due to lrauma

4. Right total retinaldetachment, Lett congenitalectopic pupil and cataract

5, Bilateral labrador keratoPathY

o, Diabetic retinoPathY

131 Chapter 3

(b\ Prevatence Rates for Binoanlar Blindness

comparison with Age specific prevalence rates for binocular blindness were calculated to allow definition previous Australian surveys. Tabte 3.3.4.4 presents rates using both the Australian

us¡ng the sample as a of blindness and w.H.o. criteria. Two sets of rates were calculated,lirst from the 1990 population denominator and second using the source population, as estimated individuals existed in the lists, as a denomínator. An assumption was made that no other blind factors: source population. The validity of the source population relied on two

(sect¡on and secrt¡on 5'3) 1 The accuracy of the population l¡sts 3.2.6

2 The ability ol the survey to identify and examine blind individuals.

health ldentification of blind individuals utilized two methods. During the survey, community not have workers with tocal knowtedge were ut¡lized to identily blind individuals who might homelands or othen¡rise attended the clinic to be examined. sometimes trips to remote Even if some outstat¡ons were undertaken to meet and examine these blind individuals. another individuals were missed on one tdp to a particular community, they often presented to

community on a subsequent lriP.

The records A second method utilized the medical records available in each community clinic. reviewed' of people aged 40 years or more, who had not been examined in the survey, were

Although two additional cases of monocular blindness were found, no binocularly blind of visual individuals were identified. The limilation of using the medical records as a check

status are:

1. Records may not have existed, particularly if the person was healthy or was

new to the communitY.

2. The case notes may not have recorded visual acuity or visualproblems' This lew was unlikely for someone who was blind and had been in the community for a

years, as there had been regular annual or 6 monthly visits to the comrrunities for the

past ten years by the Trachoma Pogram.

132 Chapter 3

mads 3. The accuracy of the record could be questioned; however, only notes by ophthalmologists, qualified doctors or hospital discharge summaries were

accepted.

poss¡ble blind individual had It was the opinion ol the investigarors that atthough it was that a prevalence blindness using data been missed, it was unlikely. Therelore estimates of the of were reasonable' from the sample as a numerator and the source population as a denominator

Table 3.3.4.4: Prevalence rates for binocular blindness by age€roup

Australlan defn.

Sample group Assuming no other blind individuals in unexamined PoPulation

per Age-group Rate per 1000 Rate 1000

20-29 21221 = 9/1000 21524 = 3.8/1000 50-59 11112 = 8.9/1000 1/118 = 5.3/1000 60+ 19/136 = 140/1000 19t179 = 106/1000 Overall 22t1514 = 14.5/1000 22J2789 = 8.0/1000

W.H.O. defn.

Sample group Assuming no other blind individuals in unexamined PoPulation

Age-group Rate per 1000 Rate per 1000

20-29 21221 = 9.0/1000 ?/524 = 3.8/1000 50-59 11112 = 8.9/1000 1/118 = 5.3/1000 60+ 17t136 = 125.0/1000 171179 = 95.0/1000 Overall 20t1514 = 13.2/1000 20t2789 = 7.U1000

133 Chapter 3

occurred ¡n the 60+ age- As expected, the highest prevalence rate for binocular blindness definition of blindness' group. Using the sample group as a denominator, and the Australian census datawere the prevalence rate was 140/1OOO in this age-group. lf projected 1986 used for the denominator the crude rates were :

proportion Rate per 1000

1 1/1000 Australian defn.: 22t2002

1 000 W.H.O. defn.: 2012002 0/1

the two The results using projected 1986 data were approximately intermediate between

previously presented rates.

134 ChaPter 3

3.3.5. Monocular Bllndness

or worse, and Monocular blindness was defined as: vision in the worse eye of cF at 1 metre vision in the better eye of 3/60 or better ( W'H.O. defn')'

proportion each sex and From a sample of 1514, 66 o¡ 4.4o/owere monocularly blind. The for age-group is outlined in Table 3.3'5.1.

Table 3.3.5.1: Monocular Blindness: Age-Group by Sex

Age' Females Males o/o Group No. MonocularlY No. Monoculafy "/o Blind Blind

0-1 0t29 0t26 2-4 1/66 1.5 0t73 5-9 0t152 0t121 1t82 1.2 1 0-14 0/1 00 15-19 0t74 0143 20-29 1/135 0.7 1/86 1.2 30-39 4t92 4.4 1/55 1.8 40-49 2t67 3.0 3/65 4.6 50-59 6/60 10.0 4t52 7.7 60+ 27t74 36.5 15t62 24.2

41t849 4.8 2sl665 3.8

to The age-standardized odds ratio of monocular blindness, when females were compared

male was: OR=1.55,95% C.l.:0.81 to 2.56, (X2=1 .46 (p-Q.227)). The causes of monocular

blindness are given ¡n the Table 3.3.5.2 (A) and (B) for females and males and respectively.

135 Chapter 3

cause. Table 3.3.5.2 (A): Causes of monocular blindness. Age-group by

Females

Cause

Total Age- Phthisical Cataract Trachoma Corneal Other Group or absent Opacity

I 0-4 1

1 0-14

1 5-19 I 20-29 1 4 30-39 3 1 2 40-49 2 2,3 I 6 50-59 2 1 3 I 4 27 60+ 4 13 6 3

41 Totals 7 18 7 5 4

1 . Dense right posterior capsule fibrosis

2. V¡treous haemorrhage

3. AmblYoPia

4. Pseudophakia(postsurgicalcomplication)

136 Chapter 3

cause. Table 3.3.5.2 (B): Causes of monocular blindness. Age-group by

Males

Cause

Total Age- Phthisical Cataract Trachoma Corneal Other Group or absent Opacity

11 I 1 0-14 15-19 2 1 I 20-29

1 I 30-39 3 40-49 2 1 3 1 4 50-59 1 2 4 5 1 15 60+ 4 5 4 I

4 25 Totals 7 7 4 3

1. AmbyloPia

2 Right traumatic mydriasis and opt¡c atropfry (trauma)

3. Retinaldetachment (trauma)

4. Complications of cataract extracrt¡on

5. Traumatic cdaract and optic atrophy (trauma)'

years. In those CI the 66 people with monocular blindness,52 (78.8%)were aged 50 or more in this age- individuats aged 60+ years 42t136 (30.9%) were monocutarly blind. lndividuals eye group tended to have poor vision in their remaining eye, whereas vision in the non'blind

was good lor individuals younger than 50 years.

137 Chapter 3

(a) Causes of monocular blindness:

Allbut three cataracl:25166 (37.g./.). The cataracts were predominantly of the senile type' years was congenital' cases occt¡rred in the 50+ age-group. One in a child aged 4

was found phthisical or Absent Globe: 14t66 (21.2o/ol. Trauma, at least as an initiating event, were not determ¡ned' to be the cause of 11 cases ol phthisical or absent eyes. Other causes and 5 cases Trauma also caused two cases of optic atrophy, one case of retinaldetachment 9f comeatopacity. Trauma made up a totalof 19/66 (28.8%l of allnronocularblindness.

in eight corneal opacity: 8/66 (12.1%). corneal Opac¡ty, not due to trachoma, was found individuals. The underlying cause was attributed to either trauma or infection; however a

precise determinat¡on of cause was otten ditficult'

in Trachoma: 11/66 (16.6%). Trachoma caused monocular blindness from c¡meal opacity paired for trachoma eleven individuals. Because there is a high correlat¡on between eyes greater dsk of sevefity, the ¡ndlviduals who were monocularly blind lrom trachoma, were at causes. binocular blindness, compared to those individuals monocularly blind from other

Tabte 3.3.5.3 identilies the individuals with rnonocular blindness at risk of binocular blindness

from trachoma.

Other: 8/66 (12.1"/"1 The categories included ambylopia, optic atrophy and iatrogenic

causes.

(bl tndividuats at risk ol binocular blindness:

were fourleen Corneal opacity or trachoma tñchiasis can lead to blindness if neglecied. There were fifteen individuals who wêre monocularly blind and had no signs of trachoma, but there visual individuals (zg.o"/"lwho were nonocularly blind and were at an immediate dsk of fudher years; impairment because of trachoma (Table 3.3.5.3). The maiorily were aged 60 or more people however one lemale was in the 30-39 age-group. There was also a group of thirteen or aged 59 or less who had trachoma trichiasis and although not yet being monocularly

138 Chapter 3

ophthalmic care were not binocularly blind, were at risk of visual impairment if adequate available (Table 3.3.7.4).

Table 3.3.5'3: lndividuals with monocular blindness, at risk of

binocular blindness from trachoma

Trachoma status in the non blind eye

Total Age- Nonnal Trachoma Trachoma Tdchiasis Group scaning or ComealOPacitY

F M F M F M

I 1-4 I 0 5-9 I 10-14 I 0 1 5-19 2 20-29 1 I

1 5 30-39 1 1 2 5 40-49 2 1 2 10 50-59 2 1 4 3 42 60+ 2 18 I 9 5

Total 77 24 13 10 5 66

19.7"/o 36.4"/" 7.6"/" 15.2/" 100/o "/" Ol 10.6% 10.6% total

139 Chapter 3

3.3.6. Vlsual AcultY

for adequate and low Visualaotity was grouped into the World Heatth Organization categories of vision. Another vision (w.H.O. 1979). Tabtes 3.3.6.1 and 3.3.6.2 outline these categoñes in both eyes)' category of visual acu¡ty is also presented for completeness (6/6 vision

(1) visualacuity Table 3.3.6.1 Age-group vs sex for 6/6 vision in both eyes and category W.H.O. defn.

6/6 Vision in both eYes Between 6/6 and 6/18 vision in both eYes o/o o/o olo Age-grouP females o/o males females males

26 100.0 0-1 29 100.0 26 100.0 29 100.0 2-4 65 99.4 73 100.0 66 100.0 73 100.0 121 100.0 5-9 152 100.0 121 100.0 152 100.0 r0-14 100 100.0 81 98.8 100 100.0 82 100.0 43 100.0 15-19 73 98.6 43 100.0 74 100.0 100.0 20-29 130 92.3 84 97.7 133 98.5 86 100.0 30-39 81 88.0 51 92.7 92 100.0 55 100.0 40-49 52 77.6 50 76.7 66 98.5 65 100.0 50-59 34 56.6 32 61.5 54 90.0 52 45 72.6 60+ 14 18.9 13 20.9 42 56.8

Total 730 86.0 574 86.3 808 95.1 648 97'4

in males The proportion of individuals with visuat acuity ol d6 in both eyes was found equally vision. There was a and females. ln the age-group 0-19 years almost every individual had 6/6 was small dectine in the proportion of individuals with 6/6 vision until the 50+ agegroup individuals in reached where less than 60% of all individuals had 6/6 vision. However, ¡f those the lhe next category of visual acuity were included (between 6/6 and 6/18 in either eye), poor in older decline was evident only in the 60+ age-group, indicating the visual status

individuals.

140 Chapter 3

vision Table 3.3.6.2: Agegroup vs sex for categories (2) and (3) of low W.H.O. defn.

Between 6124 and 6/60 vision in 3/60 vision in both eYes both eyes 6l o/o % Age-grouP female o/o male t6 female male

0 0 0-1 0 0 0 2-4 0 0 0 0 5-9 0 0 0 10-14 0 0 0 0 0 0 1 5-19 0 0 20-29 0 0 0 0 30-39 0 0 0 0 0 0 40-49 ,1 1.5 0 50-59 5 8.3 0 0 0 60+ 18 24.3 12 19.4 2 3.2 0

Total 24 2.8 12 1.8 2 0.2 0 0.0

had a greater Low vision categories (2) and (3) occurred mainly in the 60+ age€roup' Women finding prevalence of low vision in the younger age-groups, but the numbers were small' This lf the was consistent with the prevalence data lor binocular blindness (section 3.3.4). visual categories for low vision and blindness were combined the crude odds ratio of lO 3'58, impairment when women were compared to men was: oR=1.93, 95%'C.1.:1.06

odds ratio was: OR=2.02,95i'o C.1.0.92 to 1X2=+.63 (p=0.031)). ln the 60+ age-group the

4.44, (X2=3.01 (p=0.083)).

14',1 Chapter 3

3.3.7. Trachoma

presence or absence of The trachoma grading system used in this study was based on the for use by the five selecled,,key* signs. ouilined by Thylefors et at. (1987), it was devised of trachoma at the non-specialist to produce a simplified grading system for the assessment produces good observer agreement' community level. studies have shown that this system to that of the more and that the ¡nterpretation of findings recorded generally corresponds

detailed grading system (section 1.4).

lollicular Tables 3.3.7.1 - 3.3.7.5 give age-SpecifiC results for each of the live trachoma signs, (TT), comeal trachoma (TF), trachoma intense (Tl), trachoma scaning (TS), trachoma trichiasis

opacity (CO).

142 Chapter 3

Follicular trachoma:

Table 3.3.7.1: Prevalence ol trachoma follicles by age-group and sex

females males % of age- Age- No. %ofage- No % of age- Total group Group group group

10/55 18.2 0-1 8129 27.6 2t26 7.7 36.0 2-4 25/66 37.9 25t73 34.2 s0/139 3 0.4 5-9 50/1 52 32.9 331121 27.2 831273 15.9 931182 I 8.1 1 0-14 20/1 00 20.0 13182 1.7 15-19 1174 1.4 1t43 2.3 21117 20-29 7t135 5.2 5/86 5.8 121221 5.4 4.1 30-39 6192 6.5 0/55 0.0 61147 1.5 40-49 1t67 1.5 1/65 1.5 2t132 50-59 2t60 3.3 2152 3.8 41112 3.6 7.4 60+ 6174 8.1 4t62 6.5 10/r 36

Totat 1261849 14.8o/o 86/665 12.goh 21211514 14.Oo/o

(W.H.O. 1989). The Folticutar trachoma (TF) is a measure of infective disease in a community The peak prevalence occurred in lhe 2-4 age-grouP, closely lollowed by the 5-9 age'group.

lowest prevalence occurred in the 15-19 and 40-49 age-groups' The youngest age'group' of the 14 years and less, had the bulk of the disease burden, 176 out o1212 cases or 83o/"

total.

30-39 age Sex differences in the follicular trachoma rates were observed ¡n the 0-1 and The differerrce bracket, but the numbers were small and the differences were nol significant' was also in the 5-9 age-group between males and females, although having larger numbers, ratio of not significant: Mantel-Haenszel ¡2=1'o (p=0'gtz)' The age-standardized odds 0'98 to follicular trachoma for females relative to males was: M-H weighted OR=1'36, 95% C'¡' ol the 1.89, (X2=3.30 (p=g.g69)), using the five age-groups from 0-19 and the aggregate total

20+ age.grouPs.

143 Chapter 3

graphically F¡g 3'3'7'1 The relationship of follicular trachoma with age is best appreciated

Figure 3.3.7.1: lndividuals with folliclular trachoma by age-group

60

50 ø Females I Males 40

Number of 30 lndividuals 20

10

0 S1 24 $9 1G14 1$19 be 3oæ 4o4S 5o50 m+

Age-group

144 Chapter 3

Trachoma intense:

Table 3.3.7.2: Prevalence ol trachoma intense by age-group and sex

females males o/o %of ag+ Age- No. loolage' No. ol age' Total group group Group group

6/5 s r 1.0 0-1 2t29 6.9 4t26 15.4 2.9. 2-4 2t66 3.0 2t73 2.7 4/139 2.2 5-9 6t152 3.9 01121 0.0 61273 0.0 11182 0.5 1 0-14 1/1 00 1.0 0182 o 0.0 15-19 0174 0.0 0t43 0.0 1117 4.1 20-29 6/1 35 4.4 3/86 3.5 91221 101 1 47 6.8 30-39 5t92 5.4 5/55 9.1 6.8 40-49 6167 8.9 3/65 4.6 9/132 6.2 50-59 4t60 6.7 3t52 5.8 71112 8.1 60+ 4174 5.4 7t62 11.3 11/136

Total 36/8 49 4.20/o 27 1665 4.10h 63/15f 4 4.2To

15-19 age'group' The table demonstrates the low prevalence of trachoma ¡ntense (Tl) in the prevalence in the and the variable but overall low prevalence in older age-groups. The year who had younger age-groups was also low except among children aged less than ono

the peak Prevalence of 11.0%.

(in which 83% The relationship of Tl with age was the reverse of follicular trachoma (TÐ, in individuals occurred in individuals younger than 14 years), with 73% (46/63) of Tl occuning

overthe age ol 14.

145 Chapter 3

Trachoma scaning:

Table3.3.7.3:Prevalenceoftrachomascarringage{roupbysex

females males o/ool %of age Age- No. age- No %of age- Total group Group group group

0/5 5 0.0 0-1 0129 0 0t26 0.0 1.4 2-4 0/66 0 2t73 2.7 21139 4.8 5-9 5t152 3.3 8t121 6.6 131273 6.1 1 4l 182 7.7 1 0-14 9/1 00 9.0 5t82 7.0 r 6/117 1 3.7 1 5-19 13174 17.6 3143 I 20-29 231135 17.0 21t86 24.4 441221 9.9 30-39 37t92 40.2 14t55 25.5 511147 34.7 40-49 38/67 56.7 25t65 38.5 63/1 32 47.7 51 .8 50-59 35/60 58.3 23t52 44.2 581112 60+ 68174 91.9 53t62 85.5 1211136 89.0

Total 228t849 26.97o 154/665 23.2oh 38211514 25'2o/o

peak Trachoma scarring is a measure of past infection in the c¡mmunity. As expected the

prevalence occurred in the older age-groups. The age-standardized odds ratio of lrachoma to 1'79' scarring for females relative to males was: M-H weighted OR=l.15, 95% C.l.: 0.96

of 0-19 age-groups and the five age- 1X2=Z.ga3 (p=o.ogt)), using the aggregate totat the

groups in the 20-60+ age bracket. The prevalence increased w¡th age, reflecting either the recent past burden of trachoma disease or the diminishing prevalence ol active infection in

years. Figure g.g.7.2 highlights the age effect and demonstrates the steep increase of

scaning with age.

146 Chapter 3

Figure 3.3.7.2: lndividuals with trachoma scarring by agegroup

80

Females 60 ø I Mates Number of

lndividuals 40

20

n 0 Gl U $9 1ù14 itlg æ8 0æ 4eß 5o50 m+

Age-group

147 Chapter 3

Trachoma Trichiasis:

Table 3.3.7.4: Prevalence of trachoma trichiasis by age-group and sex

females males o/o %of ag+ Age' No. ol age- No. % of age- Total group Group group group

0.0 0-19 01421 0.0 0/345 0.0 0/766 1.4 20-29 2t't35 1.5 1/86 1.2 31221 1.4 30-39 2t92 2.2 0/55 0.0 21147 3.0 40-49 3167 4.5 1/65 1.5 4t192 3.6 50-59 4/60 6.6 0t52 0.0 41112 I 9.1 60+ 21t74 28.4 5162 8.1 26/135

Total 32t849 3.8oh 7/665 1.1o,h 39/1514 2.6%

Trachoma trichiasis is the consequence of years of trachoma infection with scarring and

inturning of the eYe lashes.

prevalence one in five individuats aged 60 or over had trichiasis (19.1%). The overall of ratio of trichiasis in the sample group was 2.6% (39/1514). The age-standardized odds 11'83' trichiasis for females relative to males was: M-H weighted OR=4'49, C'l': 1'81 to

(X2=12.17 (p<0.001)), using the age-groups in Table 3.3.7.4. The increased risk of trichiasis (TF in men and for women occurred despite similar prevalence rates lor active disease or Tl)

women (Tables 3.3.7.1 and 3'3.7.2)'

The occurrence of trichiasis in younger agg-groups, although small, ls a cause lor corrcem if adequale because these individuals are at risk of corneal opacity and eventual blindness

ophthalmic care is not available.

148 Chapter 3

Corneal OPacitY:

Table 3.3.7.5: Prevalence of trachoma comealopacity

bY age-grouP and sex

females males o/o o/ool age- Age- No. ol age- No age- Total ï"ol group Group group group

0.0 0-19 0t421 0.0 0/345 0.0 0/766 0.9 20-29 1/135 0.7 1/86 1.2 21221 30-39 1192 1.1 0/55 0.0 11147 o.7 40-49 1t67 1.5 1/65 1.5 21132 1.5 4.5 50-59 4t60 6.7 1t52 1.9 51112 2 3.5 60+ 20174 27.0 12162 19.4 32/136

Total27t84g3.2o/o15/6652.3"/o4211514 2.goh

as The results for corneal opacity were virtually the same as for tdchiasis' This was expected odds those with trichiasis usually have some degree of comealopacity' The age-standadized

ratio of corneal opacity when females were compared to males was: M-H weighted OR=1'&t' The c.t.: 0.77 to 3.48, 1X2=t .48 lp=Q.224)), using the age_groups in Table 3.3.7.5. was difference between the sexes was less marked, but when the outcome ol blindness

considered, women were disproportionately represented (Table 3.3.4.1 ).

The relationship of corneal opacity with age is dernonstrated in Figure 3'3'7'3 '

149 Chapter 3

Figure 3.3.7.3: lndividuals with cornealopacity by age-group

30

20 Number of ø Females lndividuals I Males

10

0 01 N $9 1011 1$19 tA 03 4S49 5059 æ+

Age-group

150 Chapter 3

lnflammatory Trachoma: and sex Table 3.3.7.6: Active lnflammatory Trachoma (TF+TI) by agegroup

females males % of age- No % of age- Total % of ag+ Age- No. group Group group group

16/55 29.1 0-1 10t29 34.5 6126 23.0 3 8.1 2-4 26/66 40.9 27t73 37.0 53/139 31 .1 5-9 52t152 36.8 331121 27.2 851273 18.7 10-14 21t100 21.0 13182 15.8 341182 '1143 2.3 21117 1.7 1 5-19 1t74 1.4 8.1 20-29 12t135 9.6 6/86 9.3 181221 11.0 30-39 1',|192 12.0 5/55 9.1 161147 40-49 7167 10.4 4/65 6.2 111132 8.3 9.8 50-59 6/60 10.0 5t52 9.6 111112 14.7 60+ 10174 13.5 10t62 17.7 20/1 36

TotaIl56/84918.9ololr0/66516.5%26611514 17.6%

Note: individuals with both TF and Tl have only been counted once

of the The addition of the trachoma intense and follicr¡lar trachoma gives an overall impression

distribution and extent of active inflammatory d¡sease in the sample.

prevalence of inlective Reflecting the results for trachoma lollicles (Table 3.3.7.1) the highest had almost I in 6 disease occurred in the younger age-groups, although the 60+ age-group age- individuals with active disease. Male to female differences occurred; however, the was; M-H weighted standardized odds ratio of infective trachoma for females relative to mates in Table 3.3.7.6. oR=1.2g, c.l.: 0.g7 lo 1.72, 1X2=Z.gZ (p=O.Og7)), us¡ng the age€roups

151 Chapter 3

Summary of the Trachoma Results:

pattem of disease across age-groups' The findings for trachoma demonstrated a changing

(TF+T|) predomlnated, affecling ln the young (o-14 age-group), active inflammalory disease contrast, was only found in 4'5% 2g.oo/o (188/649) of this group. Trachoma scarring, by

(29/649) of this grouP.

a lower prevalence ol active The adolescent and young adult years (15-29 age-group), had prevalence of trachoma scafr¡ng inftammatory trachoma (211338 ot 6.2o/ol, but a higher w¡th age (Figure (60/338 or 17.8o/"1. There was in fact a steep lncrease in trachoma scarring

3.3.7.21

was found in 9'7"h For the middle years (30-59 age-group), active inflammatory trachoma Tl now (3g/391). However, whereas TF had predominated in the younger age-groups, (approximately 68% of all infective contributed more to the total of infective disease than TF

trachoma in this grouP).

AdditiOnally' Trachoma scarring was found in 44.0o/o (1721391) of the 30-59 age-group' percentages being 2'6% lrachoma trichiasis and corneal opacity began to be found, the

(1 0/391 | and 2.0% (8/391 ) respec¡t¡vely.

greater proportions compared ln the oldesl age category (60+), both TF and Tl were found in the total prevalence of to the middle years, 7.4% (101196) and 8.1% (11/136) respect¡vely; scarring was the inflammatory trachoma was 14.7"/o (201136). The prOpOrtion of trachoma this group (66% greatest of any group (89%) and the greatest burden of TT and CO also fellin

and76o/" of the totals respectively).

by; a steep increase ln summary, the pattern of disease across age-groups was characterised

trachoma in the young, with a of lrachoma scarring w¡th age, a high prevalence of inllammatory (compared to preced¡ng age-groups) smaller but signifìcant increase in inflammatory trachoma relative proportions of in the 60+ age-group. A feature ol this increase was the reversal of the

'152 Chapter 3

corneal opacity were TF and Tl compared to the 0-14 age-group. Trachoma trichiasis and trichiasis' predominantly found in the middle (30-59) and old years (60+). For all trachoma this percentage was 77o/o was found in individuals aged 50+ years, and for corneal opacity

88.0%.

aged 60+ years' lt The burden of severe trachomatous disease appears to fall on individuals disease this was not was found that even though the young had high rates ol inflammatory opacity, as found in translated into a high proport¡on of secondary scaning, trichiasis or comeal years had less inflammatory trachoma endemic regions overseas (section 1.3.1). The middle with increaslng infective disease and atthough having an increasing proportion of scaning opacity. Only in the age, this again was not translated into high rates of trich¡asis and comeal prevalence of trichiasis 60+ age-group was there a large proportion of scarring and an high was also and corneal opacity. The burden of trachomatous disease in the 60+ age-group

reflected in the prevalence of visual disability for this group (section 3.3.3 to 3.3'6).

3.3.8. Servlce Role

Although not a part of the research, lhe service role was ¡mportant to the survey's acceptance

by the communities visited.

presbyopia The number of specfacles prescribed totalled 131. These were mainly for or

aphakia in the older individuals, although a feur were prescribed for myopia or hypermetropia.

previous It appeared that although many of the individuals had been given glasses by in the trachoma program trips, these were easily lost or broken and therefore had a shod life

bush. One suggestion to increase the lifetime ol the spectacles was the provision of

restraining straps for alt glasses. Although corrective glasses were welcomed by the older wear age-groups in the communities, young people who had poor vision otten chose not to

lhem, presumably because of a sense of shame or because of peer group pressure. At least

153 Chapter 3

probably as a two ¡nd¡viduats were seen who were monocularly blind due to amblyopia' consequence of not wearing prescribed lenses'

were made' A total of 54 Referrals lor eye surgery either to Adelaide or Alice spdngs also for people were given the opportunity to have either cataract surgery (44), or lid surgery

(4), for one person, rehabilitation' trichiasis (5), or laser treatment for diabetic retinopathy or

154 Chapter 3

3.4 Dlscusslon

3.4.1. Partlclpatlon Rates

proportion of the source ln the context ol a census type survey, participation is defined as the

population who were approached and agreed to part¡cipate in the survey. Approximately 54ï" Anangu Pitiantiatiarä of the total poputation identifiable from the 1990 population lists lor the This may be lands were examined; il projected census data were used the figure was 76%' survey for the compared to the approximate participation rate of 80.4% from the N.T.E'H.P. 100% Ap lands in 1976, using 1976 census data and the R.A.C.O. Report (1980). Although

participation in a survey is desirable if the group which participates is representative of the

population. source population the results obtained will be applicable to the larger total

ln the lggg/go survey all the communit¡es were sampled with vadable participation rales except those (42.1"/o lo 67.4%1. FemaleS were seen more oflen than men in all age-groups (76%) the 0- aged 0-4 years. Age categories with the highest participation were the 60+ and

gs (63%) age-groups. coincidentally these were lhe age-groups in which the major ocular

pathology occurred, blindness and trachoma respectively. The participation in the 60+ age-

in a group was comparable to the response (86.2%) in a prevalence survey of ocular disease

¡n other New Zealand town of peopte aged 65+ years (Martinez et al. 1982). Participation the purposes. age-groups varied lrom 41!oto 607o, which was adequate for statistical

The variation in participation reflected severalfactors including:

(a) The lime spent ¡n the commun¡t¡es. The longer tho stay the greater the

participation, but as the stay len$hened fewer individuals were seen each day (a law

of diminishing returns), this limiting the "eff¡ciency" of a prolonged visit.

(b) The perception by individuals of a need for eye care and assessment. Older

individuals with visual problems usuatly sought examination earlier than those with

155 Chapter 3

good visual acuity. Therefore young males and females in good health were less

likely to participate on the basis of individual need.

(c) The appreciation by the community and individuals of the survey's value perception inlluenced many who participated. For those without visual disability the participat¡on' of the survey benefiting the community as a whole increased

(d) The occurrence of ceremonies or business.

perception Factors that were important in lostering the communities positive of the survey were, pre-visit advertising, liaison with health service staff and community leaders, and the formation of relationships with families and individuals within the community. Often the to encouragement by one family member who had participated in the survey induced others attend for examination. The increased participation with time spent ¡n each community, was due to both the time spent in the field, and the gradualgrowth in relationships with members of the communitY.

Even with a participation rate ol76o/" potential biases could be ¡ntroduced into the results il the examined proportion of the population were systematically different from the unexamined.

individuals It was esrablished from medical records and health workerc that all binocularly blind were examined (section 3.3.4. b). Generally those not examined tended to be aged 20-39. A

group, in those examined, which showed little eye patho¡ogy and only some inflammatory

trachoma. The symptoms of trachoma did not appear to influence presenlation to the health

clinics at other times and ident¡fícat¡on of trachoma in these communities appears to ocq¡r with

opportunistic screening by health statf. lt could therefore be concluded that trachoma stalus

would not influence presentation to survey personnel and that an individuals presence in the

community was the main determining lactor for participation (see section 5.3).

previous surveys of eye health in Australia had not attempted to estimate the underlying

source population from which the sample was taken. Therefore, the prevalence rates

presented were from the study population (sample). Extrapolation from this rate to the

156 Chapter 3

proportion the population may be inaccurate because the residual or unexamined of to population may have rrþre or less disease than the sample. The 1989/90 survey attempted sourco see as many individuals as possible and to est¡mate the s¡ze of the underlying population. Also, at least for blindness, the 1989/90 survey sought to identily the visual the acuity of individuals not examined (section 3.3.4 (b) ). Desp¡tê methodologicaldifficulties, eye heatth prevalence rates presented (Table 3.3.4.4) gave a more accurale est¡mation ol the

lf a properly randomised of the total population than the sample prevatence rates themselves' prevalence data survey could have been done, the sample would have yielded more accurate lists, may be possible for the population as a whole. Wth the availability of the population this in the future.

157 Chapter 3

9.4.2. Vlsual AcultY Results

Binocular Blindness

or using the ln the 1989/90 survey 1.3!" ol the sample was binocutarly blind (w.H'o' defn'), (W'H'O' Australian definit¡on 1.5%. The source population prevalence rate was 7 -A1000 countries defn.), which was comparable with blindness rates of 1% in "typical" developing

(Thylefors 1gg0). The prevalence of blindness derived from the sample was also comparable

with the figure ot 1.4!o (visual acuity 6/60 orworse, N.T.E.H.P. defn.) forthe AbOriginal

populafion from a series surveyed by Taylor (1980). By contrast, an estirnate forthe state of (Banks Hutlon1985); New South wates (N.S.W.) was 1 .4 - 1.9t1000 or less than 0.2% and for and using population based data from a multi-stage random cluster eye health survey

people aged 50+ years in South Australia, a rate of 1.0 /1000 or 0'1oh (personal obtained communication with Dr. H Newland 1991). These rates that were similar to those

from major surveys in developed countries around the world (Sorsby 1966, Kahn et al. 1977)'

and highlight the gap between Aboriginal and white eye health in Australia'

data from A comparison of the causes of binocular blindness may be made using a sub'sel of

the N.T.E.H.p. survey. using the N.T.E.H.P. blindness delinition twenty eight people were

blind in the lggg/g0 survey. The causes of binocular blindness were in order, trachoma

(12t?lgl, cataract (11/28), absent or phthisical globe (3/28), bilateral labrador keratopathy

may be (1tzgl and retinat detachment with congenital ectopic pupil and cataract (11281. These in which compared to the results of a sub-sample of the N.T.E.H.P. survey (Taylor 19771,

again by coincidence, twenty eight individuats were binocularly blind. The causes were

(31281' cataracrt (1 tzgl, corneal scars (4/28), phthisis (3l28l,labrador keratopathy trachoma

(uzill,pterygium and cataract (1t281and interstitialkeratitis (1/28). The crude prevalence ol wers not blindness was similar , 1.8o/o vs 2.2To respectively. Age-specific prevalence rales

given in the results, nor was an age/sex distñbution recorded. However, it was established

that examination of the sample occurred on the A P lands (Personalcomrn¡nication with Prof. proportion Taylor 1g92), therefore it could be assumed that the sample included a large of the

158 Chapter 3

source population is Anangu Pitiantjatiara. The equivalence of both the sample and prevalence rates and causes of established in chapter 4 (section 4.3), comparisons of in each survey, trachoma blindness would therefore be valid. cataracts were a major cause diseases of predominated in the 1989/90 survey but not in the 1977 study. However, adding and interstit¡al kerat¡t¡s) the anterior segment (corneal scars, labrador keratopathy, trachoma had approximately together in each survey, demonstrated that these preventable diseases

equalproportions in each year (ßlze vs 10/28)'

people in and westem comparison of the results with a larger sample of Aboriginal central with the Australia reported by Taylor (1980), found consistency in the causes of blindness was due to cataract, 19gg/90 survey. Just over half the blindness in Taylo/s second study gtobe. trachoma was not as one third to corneal disease and one tweltth to loss ol the Again However' important in causing blindness in the earlier work compared to the 1989/90 survey. age, he found using a case control method that blindness was correlated with increasing

trachoma and history of trauma, findings that were consistent with the 1989190 survey.

disease of The results of the N.T.E.H.P. survey (R.A.C.O. Report 1980), demonstrated that in blind (Table the lens and cornea were the most common visual system abnormalities the

2.2.11. Trachoma atthough being important did not predominate as in the 1989/90 survey, people. A but preventable causes of the anterior segment were important for Aboriginal

further analysis of this finding occurs in the discussion of the trachoma results.

The prevalence of blindness in males and females was significantly d¡fÍerent in tho 1989190 in 1976- survey (section g.3.4). This had been found before in the "Red Centre' zone the (R'A'C'O' 1979 survey. A difference which was not significant ¡n the other twelve zones

Report 1980 P62).

blindness Bitateral bt¡ndness was mainly found in the 60+ age-group, 86% (19122) of the total

(Australian delinition). This is roughly comparable to the risk of blindness in Amedca, where

159 Chapter 3

aged less people over the age of 65 years have a 10 fold risk of blindness, compared to those than 65 years (Pizzarello 1987)

Monocular Blindness

ln the lggg/go survey the crude prevalence rate per thousand for monocular blindness

(N.T.E.H.p. defn. of 6/60 vision or less in one eye) was 56/1000 (85/1514 using data from for table 7 appendix 1). This can be compared with a crude prevalence rate ol37t1O00 the

as a ',Red Centre,, zone in the R.A.C.O. Report (1980). ll the source population was used

denominator the rate was 30/1000 for the 1989/90 survey data . When the 60+ age-group Reporl, was considered the rate for the "Red Centre" zone was 227ñ000 ¡n the R'A'C.O' 1) lor compared to the 1989/90 result of 353/lOOO (48/136 using data lrom table 6 appendix for the the sample. For the source population the rate was 26711000. A direct compañson

Anangu pitjantjatjara lands or Yalata could not be done because age-specilic data was not

available.

It was difficult to explain the apparent increase in monocular blindness (from 1976 to

lggg/go), when the sample was considered. The fall as it occurred for the source population

was more in keeping with the trend demonstraled for both poor vision (Table 4.3.4) and

binocular blindness. The high prevalence rate ¡n the 60+ age'group may have been due to a

cohort effect, w1h a group of monocularly blind individuals in 1976'1979 being counted again

ln t9g9/90. The results for monocular blindness may also rellect an increased identification ol

monocularly blind people in the 1989/90 survey compared to the N'T.E-H'P' survey, or the

differences ¡n the populat¡on distribution between the 1989/90 survey and the sample in the

'Red Centre" zone (see Table 4.3.6).

Comparison ot the causes of rnonocular blindness in the 1989/90 survey with the 1977 Taylor

study demonstrates that cataracts were the most important cause in each survey 39.4% vs

4g.6y" respectively, anterior segment disease was next 39.4o/ovs 24.3o/o and trauma as an

160 Chapter 3

28.8% underlying cause, was respons¡ble for approx¡mately a third of all monocular blindness vs 43.2o/o.

Comparison with international studies highlight factors that contribute to monocular blindness.

Firs¡y, an eye health survey of a rural Appalachian community in Kentucky (USA) gave the causes of monocular blindness as catarac"t, trauma, amblyopia, and macular degeneration. lnfectious causes were not important (Dana et al. 1990). ln the 1989190 survey cataract and

were trauma were the major causes with infection third. Geographically and racially the surveys dissimitar, but both areas were rural and socioeconomically d¡sadvantaged, with restricted access to ophthalmic services.

Secondly, an eye health prevalence survey atnong the Turkana tribe in north-west Kenya' an

area which was climatically and socioeconomically similar to the 1989/90 survey was reported

in 1gg0 (Loewenthal and Pe'er 1990). The crude prevalence of monocular blindness was

6.g%. Trauma contributed significantly to rnonocular blindness (21%l compared to the figure

for the 1989/90 survey o128.8!" (19/66). Senile catarac"t was a cause in 13.1olo (8/61)' cf

(14161)' 37.6o/o (25166), trachoma 14.8% (9/61) cf 16.6% (11/66), and comealopacity 22.9oh

precise was hard to d 12.10/o (g/66) of all monocular blindness. For cornealopacity a cause

determine because the h¡stofies were imprecise. A difficulty found in other surveys in

undeveloped regions (Potter 1991).

The compadsons are qualitative, but, the f¡rst study demonstrates that cataract can be an

¡mportant cause of blindness in developed countries, with factors such ¡solation and

economic disadvantage possibly contributing. The second study shows that desp¡te climatic

and economic similarities with Central Australia, and although the causes of monocular

blindness were the same, the relative ¡mportance of the causes may be diferent for each

region. This has implications for the prevention and treatment of bl¡ndness and trachoma

(section 6.2)

161 Ghapter 3

VisualAcuity

previous The results of the 1989/90 survey for visual acuity were similar to those from surveys conducted in Central Australia. The majority of the population (86%), had visual acuity of 6/6 in both eyes. "Good" vision (Visual acuity ol 6t12 or better in both eyes (N.T.E.H.P' defn')) per was found in gg.5% of the sample, this was comparable to the prevalence rate of 908

results a thousand found for the "Red centre" zone in the R.A.C.O. Report (1980). The of survey conducled in the 1970s, showed that visual acuity of 6/6 or better in at least one eye was found ing}o/oof the population from Yalata and 91% lrom Emabella (Eórvards et al. 1976).

Visual acuity of 6/9 in both eyes or worse was found tn 17"h of people from Yalata and 9% of

people from Ernabella, cf. 1O/olo¡ the 1989/90 survey (appendix 1 Table 7).

'162 Chapter 3

3.4.3. Trachoma Results

The five sign system simplified the trachoma grading procedure (section 1.3.1 (2.1.4)).

Although some difficulties were experienced with the current w.H.o' grading scheme' it

appears to reflect inflammatory and cicatricial trachoma accurately. lt is also subiect to less

intraobserver and inter-observer erors than previous more compl¡cated schemes (Tjelsch et (Tl) in older al. 19g7). However, some difliculties were found in grading trachoma ¡ntense

individuals. They often had scarred tarsal plates, with no normal blood vessels and irritated

eyes from smoky fires, their everted upper lids were therefore similar in appearance to

trachoma intense.

This was rellected in the results when the age distributions ol individuals with follicular

trachoma (TF) and Tl were compared. TF was found predominantly in the young; for those

aged 14 years or less the prevalence was 27!" (176/649) and for individuals more than 14

years 4.Zo/o (36/865). Of all individuals affected by TF, 83% were in the less than 14 age-

group. This finding was consistent with the literature, in which trachoma is acquired soon after

birth (section 1.3.1 (2.1.5)). lncontrast, Tl had a reverse relationshipw¡th age, with 73%

occuning in individuals aged 15 or rnore years, although the peak age-specific prevalence rate

occurred ¡n the O-t age-group 6/55 (1 1.O/ol, which was consistent with previously published

resutts. The prevatence in those aged less than 14 was 2.6% (171æ9) and in those aged

more than 14 years 5.3% (46/865). These results were unexpected and require furlher

explanation.

A high prevalence (over 15%) of severe tarsalconjunclival irflammation (Tl) in age{roups less

than 10 years of age indicates a cont¡nuing involvement of a new cohort ¡n the trachomatous

blinding process (Darougar and Jones 19S3). The low prevalence ol Tl in 1989/90 survey for

this age-group suggests they will not progress to the blinding complications of trachoma'

The differences in the age distribution of TF and Tl for the 1989/90 survey may be due to a

changing pattem of disease or systematic grading error. Other surveys have not noted such a

retationship with age, but the prevalence of the active inflammatory disease in the Anangu

163 Chapter 3

pitjantjatjara tands has declined since 1976. lt may be that susceptible individuals continue to younger suffer poor eye health due to both old scarring and neglect, whereas the age-groups have grown up with improved housing, ditferent hygiene practices and better access to trealment services. The two groups, each forming a cohort, would therefore be expected to have different patterns of disease.

Systematic grading errors might also explain the differences. Misclassification may have occurred in grading when tarsal plate scarring with diffuse fibrosis or formation of a fibrovascular membrane obscured the large deep tarsal vessels, giving the appearance of fires (which inf tammatory thickening (Thylefors et al. 1987). Eye irritation from dust and smoky the older Aboriginal peopte often spend much time around) would have conlused the clinical

assessment lurther. A study on the reliability of photographs for grading trachoma in field- was studies demonstrated problems with the photograph grader over diagnosing Tl when Tl

not observed ctinically (West and Taylor 1990). A reliance on not identifying deep tarsal

vessels when grading Tl appeared to be the cause. The same reason for misclassification of Tl

could have occurred in the 1989/90 survey'

Although this type of misclassification was not analysed, on review, the results for individual

communities showed significant variation in the prevalence of trachoma ¡ntense in the 60+

age{roup. This variation could not be accounted for by the presence of ditferent graders and

therefore the high prevalence of TF in the 60+ age-group may be due to outbreaks of active

trachoma in different communities during the survey, with lndividuals most susceptible

because ol previous scarring demonstrating signs of active disease.

The prevalence of trachoma scarring reflects past d¡sease in the community (W.H.O. 1989).

The almost exponential relationship with age demonstrates this well (Figure 3.3.7.21.

Although scarring may lake years to develop, other surveys in endemic regions have found

that scarring can occur early, predisposing the individualto repeated infection and scarring

19 over many years (section 1 .3.1(2.1 .5)). The finding that only 5.9V" øínn of those aged

years or less have any scaning (and most ol it mild), reflecls both the low prevalence of TF and

164 Chapter 3

Tl, and the probabte influence of treatment programs. The higher prevalence in age-groups poorer above 20 probabty reflects infection prior to 1976 when living conditions were and treatment Programs unavailable.

The results for trachoma tdchiasis (TT) and cornealopacity (CO) were very similar' and again period 1976 reflect many years of severe disease. Prevalence rates had declined in the from

to lggg/g0 (section 4.4). A cohort etfect may have occurred, with a group of individuals aged

in their 40s, 50s and 60s in 1976 who had severe disease, ageing by 15 years and being re-

counted in the 19gg/go survey. Although this cannot be proven, it could explain the

persistence of severe disease despite the diminishing prevalence of act¡ve inllammatory

disease

Some individualdata sheets are stillin existence in Emabella (as a collected set) and Yalata

(incorporated into the medical case notes of the clinic). A retrospectlve longitudinal study

would give a more accurate picture but follow-up of individuals would be difficr¡lt.

There is some evidence of severe disease in younger individuals, 33% percent of tñchiasis

occurs ¡n the age-group less than 60 and 241" ol the corneal opaoity. However, when

comparisons were made with surveys overseas, the absence of TT and CO in the < 19 age-

group indicate a patlern of less severe disease (Jones 1975).

Two other results were important. The increased prevalence of trachoma tdchiasis in females,

w¡th an odds ratio of 4.49 compared 1o males (across all age-groups), has important

implications. Males and females had non-significant differences in prevalence rates for active

inllammatory disease and tarsal scarring, so why did women have more trich¡as¡s? The

increased risk of trichiasis for women has been observed in other surveys (section 1.3.1

(2.1.7)). A recent survey in the Nile delta lound 75% ol older women had trichiasis compared

lo 57"/o of males. Severs conjunctival scarring was more common (84olo vs 58% respectively)'

tnflammatory disease was found to be equally distributed between the sexes (Courtright et al.

1 98s).

165 Chapter 3

plausible young women Several explanafions have been offered in the past. The nþst is that acquiring have greater contact with young children who have high infectivity rates, thereby for a higher trachoma more often (section 1.3.1 (2.1.7 b)), an explanation which accounted prevalence of cicatricial trachoma in girls compared to boys (10% vs 6%) in a study from

Mexico. There was also a higher prevalence of active trachoma in girls (29% vs 19% respectively) (Wilson et al' 1987).

There are two other explanations of why trachoma is more severe in women than men. programs. Women may have poorer eye hygiene, or may have decreased access to treatment

There is no supporting evidence concerning eye hygiene, nor has differential access to

treatment been identified in trachoma studies as a cause for a higher prevalerrce rates;

however, the public health literature has looked at the social determinants ol health and

disease. Among the many faclors known to have have some etlect on health status, access

to medicalcare for disadvantaged groups, eg. migrants, unemployed, the poor, can add to the

burden ol disease (Syme 1986). Women were found to have more binocular blindness than The men (17lg49 vs 5/665) with the causes being trachoma and cataracl, 1x?=t.ot P=0.045).

results for monocular blindness also appear to support this, with 24.4o/o (10/41) of nonocularly (5/25) blind temales being at risk of binocular blindness due to trachoma compared 20.09/o with the mates oR=1.29 ,95% C.t.: 0.34 to 5.53 (Table 3.3.5.3). This finding is in agreement

Nepal Btindness Survey (Briltiant and Brilliant 1985)where allthe people bilaterally blind from

trachoma were wome n (2.4!" of all blindness) and a survey in the Transvaal region ol South

Africa where women were more likely to be blind or have impaired vision from lrachoma than

men (Ballard et al. 1983).

The hypothesis of differential access to treatment appears to be supported by data for the

prevalence of sight-limiting cataracÍt in males and females. Case contro! studies examining

age-related cataract (corticalcataract) have determined the following odds ratios (female dsk of

cataract compared to male risk), OR=2.20,95/o C.t.:1.56 to 3.12 (The ltalian American

cataract study Group 1991), oR=1 .51 , 95o/o c.l.: 1 .07 to 2.12 (Leske et al. 1991). ln the

166 Chapter 3

1gg9/go survey monocular blindness due to cataract was found in 28"/" (71251 of males compared lo 44"/" (18t41) of females oR=2.01,957o C.l.:0.61 to 6.75, but again the difference was not significant at the p<0.05 level'

is an Why women are at grealer risk of calaracts or are less likely to access ophthalmic care

important question. Socio-cultural processes may be the determining factor. Two studies physician than have identified that in traditional societies men are more likely to use services

women (Lieban i97g, Morsey 1980). A study of those blind from cataract ¡n the Nepal

Blindness survey found that men were nrcre readily provided with cataracÍ surgery regardless to of wealth or access lo care, whereas women received surgery if barriers such as distance

treatment and cost were minimat (Brilliant and Brilliant 1985). Another study ¡n the Northem

Transvaal (Bucher and ljsselmuiden 1988), found that women were 1'6 times less likely to prevalence have undergone cataract surgery than men. However, a sex difference lound in a

survey of blindness in a ruralAppalachian community could not be ascribed to the prevalence

of cataract or the use of surgical services (Dana et al' 1990); and in a survey of blindness in

Saudi Arabia there was a strong indication that women formed cataracts earlier than men and

had more procedures for cataract removal (Tabbara and Ross-Degan 1986).

ln conclusion, it appears that data from the 1989/90 survey and the literature supports precise differences in the prevalence ol trachoma between females and males. Although a

reason is unknown, the data from the survey supports differential access to treatment as

being a possible factor. The answer is not clear and willonly be determined by further study.

'167 Chapter 4

chapter 4. changes ln the Prevalence of Trachoma and Blindness in the Anangu Pitlantlatlara: A Quantltatlve Comparison Between the f 989/90 Survey and Prevlous Surveys

168 Chapter 4

4.1 ¡ntroductlon

used make comparisons' a Although the qualitative descriptions of past surveys can be to have been used quantitative analysis would be more accurate. cross-sectional compafisons review in 1985 (T'E'H'P' before to quantitate changes over t¡me. Firstly, the N'T'E'H'P' Aboriginal commun¡l¡es Report 1985), compared the resutts of trachoma screen¡ngs in twenty N.T'E.H.P. survey' Wh¡lst with the results from the same communities during the 1976-1979 definition, otten only the methodology of this comparison allowed for diflerences in disease review), leading to reduced small numbers were sampled in the second survey (the 1985 trachoma surveys in the conlidence in the differences observed. Secondly, a review of was shown to have declined in Northern Territory was undertaken in the late 1980s, trachoma by Meredith utilized data coastal areas but not in the inland (Meredith et al. 1989). The study may be from many surveys which were not standardized for methodology, and therefore out this ¡naccurate, although the statistical methods used appear to have smoothed ol these inaccuracy allowing crude comparisons and deductions to be made' Neither

comparisons had anempted to assess changes in visual aolity.

4.2 Methods

(a) The HyPothes¡s

To test the hypothes¡s:

"The age-specific prevalence rates of blindness and trachoma in the Anangu pìi¡".tj'ãt¡.. of South Australia have shown a statistically significant decline s¡nce the N.T.E.H.P. survey of 1976".

had to be found' Three Data which had been collected in a similar way to the 1989/90 survey 1976'1979 survey called the sets of data were identified. Original reports from the N.T.E.H.P. data lrom interim reports, data for the "Red centre" zone in the R.A.C.O. Report 1980, and

the 1985 N.T.E.H.P. review (T'E.H.P' Report 1985)'

169 Chapter 4

and the Three major ditficulties arose in the comparison between the 1989/90 survey data lggo R.A.C.O. Report. Firsily, no community-specific data were available from the 1980 report. lnterim reports compiled during each community visit in 1976 were; however, obtained from Professor Fred Hollows in sydney. unfortunately reports were not found for the ¡nterim reports communities of Ernabella and Fregon. The information contained in the included:

1. The age d¡stribution of individuals with poor vision (defined by the survey as visual in the other acuity in the better eye of 6/18 or 6124 o¡ 6/36, with similar or worse vision

eye).

2. The number of blind individuals (delined as visual aority of 6/60 or worse in the better

eye).

3. The age distribution of individuals with trachoma lollicles.

the Secondty, comprehensive results were available from the R.A.C.O. Report ol 1980' but were data for the communities in the N.w. of south Australia, the Anangu Pitiantjatiara lands, the "Red þoted with other communit¡es from Central Australia to form a distinct zone called

centre*, one of thirteen such zones that were created ¡n the report to aggregate data based

on geographical and demographic features. The communities from the Anangu Pitjantjatiara

lands made up 16% (1011 individuals) of the "Red Centre" zone total population of 6321

individuals (Appendix 2 R.A.C.O. Report 1980)'

It could be assumed that the averaged results lor the "Red centre" zone were an accurate

estimate lor the results of the Anangu Pitjantjatjara communities. Some dillerences would be

expected but the general characterist¡cs of the commun¡t¡es throughout the 'Red Centre'

zone including environmental, socio-demographic and cultural factors were relatively

homogeneous (R.A.C.O. Report 1980). In contrast, the comnu¡nity of Yalata was pooled with

37 other country towns and Aboriginal reserves to form the "Arid Eastern" zone. The

characteristics of the communities in this region were less homogeneous and the Yalata

170 Chapter 4

1980 population comprised only 2.4.o/" of the total sample. comparisons between the be inaccurate. R.A.C.O. Report pooled data and the 1989/90 survey results could therefore

Report' ln 1985, Another source of inaccuracy was the actual data contained in the R.A.C.O' Aboriginal a review of the Trachoma Program was undertaken, with approximately 2000

peopte being examined (T.E.H.P. Report 1985). lt was established that certa¡n discrepancies T.E.H.P. Report and inconsistencies existed in the 1980 R.A.C.O. Report (section 6.2.3.1 instance, zone 1gg5), both in the report itself and between the data base and the report. For with the zone tall¡es tall¡es for lollicutar trachoma in the text of the report differed inconsistently of in appendix 2 ol the report. A suspected sourcs of discrepancy was the addition the data rescreening data (in most cases lollowing a treatment program) to the tally for

represented by tables in the main text.

Although this may affect the interpretation of some results based on the "Red cenlre'zone

data (section 4.3.3), the comparisons based on the interim reports and 1985 N.T'E'H'P' data review ars not affected because the ¡nterim and 1985 review reports used frst screening

to calculate their results.

1989\90 ln summary the three sources of 1976 survey data available lor comparison with the

survey results were:

1. The lnterim RePorts

These unpublished reports provided 1976 data for lhe communities of Yalata, Amala,

same tndulkana, Mimili and Pipalyatjara. These data were compared with those for the prevalence of communities in the 1989/90 survey. The data available included the

follicular trachoma, the prevalence of poor vision by age and the total number of blind

people.

171 Chapter 4

2. The "Red Centre' Data from the 1980 R'A'C'O' Report

The results lor the "Red Centre" zone in the 1980 R'A.C'O' Report were compared lndulkana' with the 1989/90 pooled results for the communities of Ernabella, Amata,

P¡t¡antjatjara Mimili, Pipalyatjara, Kalka and Fregon, communities that are in the Anangu

lands.

3. The Results of the Trachoma and Eye Health Program Review of 1985

Yalata, Prevalence data for act¡ve inflammatory trachoma in the commun¡t¡es of

tndulkana, Amata and Fregon for the years 1976 and 1985 were available' R'A'C'O' Descriptions used in a grading manualfrom the 1976 survey and the 1980

Report were applied to the 1990 five sign system. ln some cases the same

photographs had been used to illustrate grades of trachoma, facilitating the cross-

match¡ng of grading.

(b) Comparison of Trachoma Grading and Diagnostic Cñteda

years since the The system for trachoma grading has evolved and changed in the fourteen grading 1976 survey. lt has been simplified to incorporate fewer categories and to avoid pits, particular signs by their severity. Also, previously well known signs such as Herberts papillae and pannus have been deleted from the new grading system' Therefore, direct clearly comparison ol results between surveys spanning these years is difficult. However, made' defined criteria are available for each category enabling some comparisons to be

appeared in The diagnostic criteria are outlined below. Abbreviations used are the sams as on the 19g5 Trachoma and Eye Health Program Report. The criteria for 1989/90 are based inlense are the l¡ve sign system. The diagnostic criteda for trachoma follicles and trachoma Report 1985)' defined using the description found in the 1985 N.T.E.H.P. review (T.E.H.P'

172 Chapter 4

Diag rn stic C riteria Ío r t racho m a

New diagnostic criteria lor folliq¡lar trachoma were devised by the investigators ¡n ths

1985 N.T.E.H.P review for the purpose of analysis.

The following abbreviations were used (T.E.H.P. Report 1985):

absent FOL = Follicles O=sign

1=sign just present LF = Limbalfollicles

2=sign obviously present PAP = Papillae

grossly present HP = Heôeds pits 3=sign

SC = Scaning

PAN = Pannus

TR = Trichiasis

Example

"FOL=2+" means "Follicles grade 2 or rþre'

173 Chapter 4

Folllcula¡ Trachoma

1980 Criteria

pgy¡p = 2 + PAP=1+(sameeye) or (1) FOULF = 1 combined with

or HP =2+ (same eye) LF or (2) FOL = 1 combined with =1(sameeye)

1985 Criteria

As above, Plus two new categories

FOULF = 1 ¡n both eyes

FOULF = 1 combined with SC = 1 + (same or other eye)

1990 Criteria

of LF, HP, SC, Trachomafollicles (TF ) FOL = 1+ (i.e. grades 1,2o¡ 3)with any comb¡nation

PAN or TR in one or both eYes.

described It was apparent that the 1999/90 criteria were broader, encompassing allthe criteria

for i9g0 and 19gS and including infec{ion in only one eye. Although limbalfollicles, Herberts graded TF pits or pannus were not specifically looked for, if follicles were present, it was as 1980 or Therefore, the 19g9190 criteria could overestimate TF in the community relative to the

1985 criteria.

174 Chapter 4

Severe Folllcular Trachoma

1980 Criteria

Usirq the same abbreviations as above

Severe folliq¡lar trachoma (Severity A and B combined )

FOULF = 3

p¡p=1+ or (1) FOULF = 2 Combined with (same eye)

p¡p=2+ or (2) FOULF = I combined with (same eye)

1985 Criteria

Limbal follicles were excluded.

FOL=3

or FOL = 2 combined with PAP = 1+ (same eye )

or FOL = 1 combined with PAP = 2+ (same eye )

plus two new categories:

FOL = 2 combined with PAN = 2+ ( same eye )

or SC = 2+ (same eye ) (same eYe )

orTR = 1+ (same eye )

or FOL = 1 combined with PAP = 1 and PAN = 2+ (same eye)

2+ or SC = 2+ ( same eye )

orTR = 1+ ( sarne eye )

175 Chapter 4

1989/90 Criteria

grading is The 1990 survey had no equivalent lor severe follicular trachoma as the of sevedty intense avoided in the Five Sign System. Some elements were equivalent to trachoma

discussed below.

Trachoma lntense

Trachoma intense (Tt )was a 1989/90 grading sign, equivalent to papillae grade 3 in the 1980

grading system. lt was possible to score grade 2 papillae, especially with trachoma scaning, as

trachoma intense in the 1999190 survey (usually in older individuals). Also individuals scored

as Tl in 19g9/go could have also been scored as trachoma fotlicles ( TF ) in the live sign

system. To avoid double counting, these individuals were designated as having Tl only, they

were not included in the TF data.

Therefore trachoma intense overlapped with trachoma follicles and papillae (with the addition prevalence of any combination of LF HP SC PAN or TR), and a comparison between the of

trachoma follictes in the 1gg0 Report, the 1985 review and the 1989/90 Survey requires the

grades 1, addition of Tl to the TF data. Note: (The data presented in the interim reports for TF

2 and 3 could be directly compared to the TF results for the 1989/90 survey because the

definitions were the same.).

Summary

The grading systems for the 1980 report, 1985 review and 1989/90 survey are differer¡t. An

evaluation of the diagnostic criteria was made allowing comparison of active inflammatory

trachoma between the surveys (TF + Tl in 1989/90 survey). The broader definition in the

more recent five sign system probabty overest¡mates the prevalence of trachoma in a

community relative to the 1980 and 1985 definitions.

176 Chapter 4

4.3 comparlson wlth the 1989/90 survey uslng three data sets

Three sets of data were used to make comparisons between previous surveys and the populations are 1gg9/g0 survey (section 4.2). For each data set the sample and source Crude and compared between surveys to ensure that they were approximately equivalent. either X2 age-standardized prevalence rates are then compared. The results are tested using

stratified analysis, or for two sets of data, logistic regression analysis.

(1) lnterlm Reports

(a) comparison ol the poputat¡ons lor the interim reports 1976 and tha 1989/90 Suruey.

Table 4.3.1. compares the source populations lor the communities surveyed for the

1976 interim reports with the sams communities in the 1989/90 survey' The figures

were derived from 1976 census data, 1986 census data, and 1989/90 population

l¡sts.

Table 4.3.2. compares the sample, in the 1976 interim reports and the 1989/90

survey

lnterim report data was available lor the communities of Yalata, Amata, lndulkana, Mimili

and pipatyatiara.( Note: Kalka was not otficially recognised in either the interim reports

or the 1976 census.)

177 Chapter 4

(1976 data) Table 4.3.i.: Compadson of the source populations: 1976 inlerim reports census

and 1999/90 survey (1986 census data and 1989/90 population list data)

o/o o/o 1989/90 o/o Ol Age-Group 1 976 ol 1 986 ol (Years) Census Total Census Total Populations Total Data Data List

192 11.4 <2 41 ) 13.4 130 11.2

2-5 100 ) 13.8 6-1 0 140 13.3 120 10.3 232 11-20 234 22.2 266 22.8 379 22.6 21-30 173 16.4 268 23.0 310 18.5 31-40 160 15.2 't28 11.0 206 12.3 41-50 90 8.5 89 7.6 138 8.2 51-60 60 5.7 65 5.8 107 6.4 60+ 54 5.1 99 8.5 115 6.8

Total I 052 1OO"/o 1165 l0Ûoh 1679 1000/6

Each set of data had been comp¡led in a different manner. The 1976 census data were

collected us¡ng 4 coilector distdcts and an unknown number of collectors. The collec'tors point enumerated the number of individuats present in the communities at a defined in time.

ln i976 the communities were more clearly delined than in subsequenl years and there wero

few if any homelands.

The 19g6 census was more dilficult because the populat¡on was scattered across

commun¡t¡es and homelands. Additionally, the collection of data was ¡ntemjpted for several

weeks in one area because of a dispute between a collector and a community (Personal

communication with M. Stratton of the Australian Bureau of Statistics). The 1989/90

population list by comparison was compiled over several months by an examination of medical

records, school lists, pension records, social security benefits and local knowledge. lt had

been updated Prior to its use.

178 Chapter 4

population were whatever the limitations set by the collection of data, the three distributions the populations were the best est¡mates available. comparison by age-group indicated that very similar in structure over the years.

survey Table 4.3.2.: comparison of the samples: 1976 interim reports with the r989/90

o/o ol Age-Group 1 976 l" ol 1989/90 Total (Years) survey Total survey

<2 36 5.2 32 3.5 10.4 2-5 67 9.6 96 17.1 6-1 0 135 19.3 158 11-20 155 22.2 175 18.9 21-30 89 12.8 124 13.4 10.4 31-40 72 10.3 96 41-50 63 9.0 90 9.7 6.8 51-60 23 3.3 63 60+ 58 8.3 91 9.8

Total 698 100o/o 925 1OO7o

was Although the lggglg0 population was larger than the 1976 sample the overall structure

similar across the age-groups l¡sted. There were more people examined in the 1989/90

survey in the 5i-60 age-group, approximately double the proportion seen in 1976. However,

this was the only age-group in which a ¡arge difference was found between 1976 and

1989t90. The proportion sampted in this age-group was 38.3% (23160) and 58.9% (63/107)

respectively. A difference that was s¡gn¡ficant (X2=6'5 (p=O'01))'

'179 ChaPter 4

(b,) comparison of the interim reports with the 1989ß0 suruey: Follicular trachoma

with 1989/90 The follicular trachoma prevalence for the interim reports was compared the survey (Table 4.3.3.).

Table 4.3.3.: Comparison of the folliqrlar trachoma (TF)

prevalence by age-groups: The intefim reports and 1989/190 survey

1 976 1989/90

o/o loll.1 Total o/" foll.l Total trach seen of age trach. seen of age- Age- group group group

18.7 Less 17 36 47.2 6 32 than 2 2-5 47 67 70.1 27 96 28.1 21.5 6-1 0 68 135 50.4 34 158 11-20 34 155 22.0 14 175 8.0 124 4.0 21-30 1 89 1.1 5 9.4 31-40 1 72 1.4 9 96 41-50 2 63 3.8 6 90 6.6 51-60 0 23 0.0 6 63 9.8 61+ 0 58 0.0 9 91 9.9

Total 17O 698 24o/o I I 6 925 12.5o/o

1. Trachomafollicles (TF), graded 1,2or 3 (sec{ion 4'3'2'l

The intedm reports gave results for trachoma lollicles graded as 1, 2 or 3, therelore the ir¡terim

report resutts could be directly compared with the TF resutts of the 1989/90 survey. By which contrast, in the R.A.G.O. Report of 1980, the results were given in broader categories

grouped various diagnostic features. Therefore a compadson based on these results

required grouping of the categories TF and Tl in the 1989/90 survey (Table 4'3'7'l'

180 Chapter 4

in 1976 to There was a decline in the crude prevalence rate of follicular trachoma lrom 24i"

12.s%in 1g90. The difference was statistically significant. The age-standardized odds ratio of follicular trachoma when 1976 was compared to 1989/90 was: M-H weighted OR=2.34,

95% C.l.: 1.77 lo g.20, (X2= 35.54 (p< 0.001)), using the four age-groups in the 0'20 age

bracket and the aggregated results for the 21-61+ age-group'

prevalence rate for The difference was greatest in individuals aged 0-20 years, where the in the each of the two surveys was 42.O/o and 23.6% respeclively. There was an increase

prevalence rate of TF in the 20+ age-group from 1.3% (4/305) lo 7 '5/" (35/464) which was a

signif icant increase 1X2 =1 t.82 (p

(c) Comparison of the ¡nterim reports with the 1989/90 suruey : Poor Vision

The def¡n¡tion of "Poor Vision" was taken from the 1980 R.A.C.O. Report and applied to the

19g9/go survey data. That is: Persons whose vision in their better eye was 6/18 or 6124 or

6/36. The vision in their other eye might be similar or worse. Table 4.3.4. compares'Poor

Vision" between the surveYs.

181 Chapter 4

Table 4.3.4.: Comparison of poor vision

by age-groups, for the interim reports with the 1989190 surveys

1 976 1989/90

ol o/o Poor Total fo Poor Total Vision Seen of age- Vision Seen of age- group Age-group group

0.0 Less 0 36 0.0 0 32 than 2 0.0 2-5 0 67 0.0 0 96 0.0 6-10 0 135 0.0 0 158 0.0 11-20 4 155 2.6 0 175 2.4 21-30 6 89 6.7 3 124 2.1 31-40 8 72 11.1 2 96 41-50 16 63 25.4 5 90 5.6 51-60 10 23 44.0 7 63 11.1 61+ 36 58 62.1 52 91 57.1

Toral I0 698 11.5 69 925 7.5

interim poor vision was the only measure of blindness which had age-specific data in the 1976

reports. Table 4.3.4. shows a decline in the individual age-specific prevalence rates, with the

smallest change occurring in lhe 61+ age-group. After grouping some of the age categories

the significance of this reduction was tested.

182 Chapter 4

P value Age- 1976 1989/91 Odds f ü.t.t Group RACO rePort survey Rat¡o corrected)

0.002 0-40 1 8/554 5/681 4.54 9.24 41-50 16/63 5/90 5.79 10.70 0.001 51-60 10t23 7t63 6.15 9.18 0.002 61+ 35/58 52191 1.14 0.05 0.828 Overall 80/698 69/925 1.61 7.71 0.001 (crude)

The age-standardized odds ratio of poor vision when 1976 was compared to 1989190 was: M- in H weighted OR=2.86, 95% C.l.: 1.86 to 4.75, (X2=23.2 (p< O.OOl))' using the agegroups the 21-61+age bracket and the aggregated results for the 0-20 age-group.

lf the source population was used as a denominator, the results did change (assuming no

other people with poor vision in the unsampled population).

P value Age 1 976 1989/91 Odds f ryart Group RACO report survey Ratio corrected)

0-40 18/848 5/1319 5.70 13.33 0.000 41-50 16/90 5/1 38 5.75 11.40 0.001 51-60 10/60 7t107 2.86 4.28 0.070 61+ 35/541 52t115 2.42 4.89 0.026 Overall 80/1 052 69/1 679 1.92 14.65 0.001 (crude)

interim reports gave 1 The 1976 Census gave a total in the 60+ age-group of 54 but the 58 as the number seen.

Two changes ocÆurred, the ditference in the 51-60 age{roup was no longer significant at the

0.0S level and the difference for the 61+ age-group became significant. This demonstrates

the impqrtance of defining the denominator accurately.

183 Chapter 4

(27t1000) in the 1976 interim Binocular blindness was reported as occurring in 19 of the 698 The ditference was reports, compared wirh 21 out o1925 (or 2311000) in the 1989/90 survey. population was used as a not statistically significant, (x2 = 0'18 (p = 0'675))' lf the source (p=o'311¡¡' denominator the vatues became 19/1052 and 2111679 respectively, 1X2=t.02 zon9, when a This result should be compared with the results found for the "Red centre"

significant reduction was lound for the source population estimates.

(2) "Red Centre" zone data

N.T.E.H'P Report (a) comparison of the poputations for the "Red centre' zone data in the

1980 and the 1989/90 suruey.

Table 4.3.5. compares the source population lor the Anangu Pitiantiatiara lands in

1976 (when the data for the 1980 report was collecled) and 1989/90 using 1976

census data, 1986 census data and the 1989/90 population lists.

Table 4.3.6. compares the samples for the two surveys. The age structure of the

,,Red Centre, zone was projected from the sample total for lhe communities of the

Anangu Pitjantjatiara lands as recorded in the interim reports'

The communities included in the "Red Centre' zone were Amata, Ernabella,

lndulkana, Mimili, Kalka, Pipalyatiara and Fregon.

184 ChaPter 4

"Red centre' Table 4.3.5.: Comparison of the source populations: N.T.E.H.P Report 1980

zone (1976 census data) with the 1989/90 survey (1986 census data and 1989ß0

PoPulat¡on list data)

o/o o/o o/" 1989/90 Ol Age-group 1 976 ol 1 986 ol Total (Years) Census Total Census Total Population Data Data List

0-4 156 12.4 190 13.3 277 12.4 15.5 5-9 205 16.3 145 10.1 347 466 20.9 1 0-19 277 22.0 360 25.2 20-29 198 15.5 275 19.2 422 18.9 30-39 174 13.9 132 9.2 236 10.6 40-49 122 9.7 123 8.6 189 8.5 50-59 58 4.6 88 6.2 151 6.8 60+ 78 6.2 116 8.1 147 6.6

Total r 265 l OOo/o 1 429 I 00% 2235 1000,6

Comparison between the three source populations in 1976, 1986 and 1989/90, indicated

that the age structure had not changed. The source population for the original 'Red Centre"

zone of the 19g0 report could not be calculated because the data were not readily availablo.

The total derived is for the A P lands only, whereas the distribution of the sample was taken .Red from the overall Centre" zone which included towns or commun¡t¡es such as Yuendemu,

Napperby, Santa Terasa etc. (R'A.C.O' Report 1980' appendix 2 pl86)'

185 Chapter 4

"Red zone (1976 Table 4.3.6.: comparison of the samples: R.A.C.O. Report 1980 centre"

data)with the 1989/90 survey

o/o 1989/90 olo ol Age-group 1 976 ol Total (Years) "Red Centre" Total Survey Sample Sample

31.4 0-4 2944 36.8 162 5-9 221 18.6 10-19 1 997 25.0 227 20-29 929 11.6 177 14.5 9.6 30-39 614 7.7 117 40-49 486 6.1 104 8.5 50-59 335 4.2 91 7.5 60+ 692 8.6 120 9.8

Total 7997 l OOo/o I 219 1 00%

Of all the population compadsons, the 'Red Centre" zone 1976 sample was lhs least population (Table comparable with the 1989/90 survey, despite having a similar source

4.g.S.). lt should be noted that the population totalfor the 'Red Cer¡tre" above was 7997, this 4.2. was taken from the text of the 1gg0 report, whereas the figure previously given in sect¡on

(6321) was taken from the appendix of that report. A possible explanation was g¡ven ¡n the

1985 T.E.H.p. Report. The authors thought that the higher percentage of individuals aged

tess than 20 and a proportionately lower percentage of people ¡n the greate:'than 20 age-

group, when compared with the 1985 sample, was accounted for by the rescreening of

people after treatment (ie. these individuals had been double counted). Those rescreened

were inctuded in the results for the 1980 R.A.C.O. Report. ln the Anangu Pitiantjatiara lands,

this second screening occurred in November 1976. ln the communities of Amala, lndulkana This and Fregon only individuals under the age of 20 were included (T'E.H'P. Repod 1985).

probably occurred for two reasons: the people treated tended to be younger (see the results

186 Chapter 4

school age population lor trachoma follicles in 1976, Table 4.4.6) and when rescreened, the 1977 and included was often the easiest to find and examine. Yalata was rescreened in June whether this data was individuals aged 20+. lt was not clear f rom the 1980 R.A.C.O. Report included in the linal result. on the 1976 interim When the sample was compared for the community specific data based population) samples resembled reports (which separated out the initial and rescreening the one another more closely (seclion 4.3 (1) )'

for the 'Red (b) comparison of the Trachoma Fotticte Prevalence Rates, by age-group,

Centre' zone with the 1989/90 suruey'

value in Although the "Red centre" data had shortcomings (section 4'2), they were of were indicating the trend of trachoma prevalence. The results lor poor vision and blindness in A P lands only marginally affected by including rescreening data, those re-screened the

were aged less then 20 years, an age-group that had low rates of blindness' However' by lor increasing the denominator (with people who were not blind) the prevalence rates

bl¡ndness and poor vision would be decreased.

lnflammatory Trachoma

the The prevalence of follicular trachoma (TF), from the R.A.C.O. Report was compared to was inftammatory trachoma prevalence rate (TF+T|) for the 1989/90 survey. A comparison years or less' done using the results for the "Red Centre" zone. Data for individuals aged 19

in 1976 and the 1989/90 surveys were used (Table 4'3'7)'

187 Chapter 4

for infective lrachoma' Table 4.3.7.: lnflammatory trachoma - compadng the prevalence rates - "Red zone by age-group lor the N.T.E.H.P.Report 1980 and 1989/90 Survey centre'

Age-GrouPs G1234t.è78€1G1112.131+151è171&19Totd

1976

36 31 æ30 TF 12. 318 41 337 300 n7 160 88

4É81 I\bTF 378 ,180 676 679 721 648 492 350 2ß æ

14o/o 1?/o 41o/o o/allF 3?P/" 66"/o 59:/" 4V/o 41o/" 36"/o 3?/" 8%

1989/90

TF 1061 æt73 37t83 32Æ 21ln 19/64 7tæ 3,50 or27 ?Jæ 159,610

'ft 661 4t73 1/81Í1 3,99 vn 0/64 1Æ0 0Æ0 0127 0ß6 16/610 ?tæ 165,610 ruTF 1661 32113 38/83 35p9 øn 19/64 8Æ0 3/80 0127 olo 31Vo 4o/o æh ßT" æ/" 3ú/o 1ú/o æ/" ú/" 5% æh

Odds Ratb and x2

189 Odds 1.04 237 173 180 118 137 253 8.æ 231 Ró

f Cror O.OO 11.1 5.01 GSO 4s9 092 5.04 16.6 3.3 0:14 4.6 @nneded)

p= 976 .OO1 .O?5 O11 .æ6 3g7 .v27 .OOO .0681 ß2 'ooo

TF trachoma follicles 1 . As cell value < 5 Fisher exact 1 tailed tesl used = P = 0'02

2. As cell valus < 5 Fisher exact 1 tailed test used Tl= trachoma intense P = 0'18

comparison between the two surveys showed an overall reduct¡on in the prevalence of

follicular trachoma in the less than 20 age-group from 41o/o in 1976' to 29% in 1990'

188 Chapter 4

signilicant (p < 0.05) Reductions occurred in all age-groups. The differences were statistically lor all ages except the 0-1 , 10-1 1 and the 1 6-19 agegroups'

result encompassed a It shoutd be reiterated that the addition of TF and Tl for the 1989/90 greater range of disease than the 1976 definilion of TF, consequently the 1989/90 result (section 4.2)' would be expected to overestimale inflammatory trachoma relative to 1976

presented above' Using logistic regression analysis a model was derived to fit the data

Folliculartrachoma was found to vary by age{roup and year'

The nrodel had the following estimates.

Estimate SE t stat¡stic

Constant -2.239260 0.1218586 -18.376 6.6471 Group 1 0.878363 0.1321429 Group 2 2.1 931 73 0.1 209089 1 8.1 39 Group 3 1.945486 0.1102825 17.641 Group 4 1.549272 0.1 091 487 14.194 Group 5 1.222579 0. I 092529 1f .190 Group 6 1.030494 0.1132677 9.0979

Group 7 1 .1 33989 0.1279830 8.8605 Group I 0.0 0.0 0.0 7.01',12 1 976 0.6866309 0.0979333

13, 17)' Where group represented One Of 8 age categories (average age = 1, 3, 5, 7, 9' 11,

1976 compared A decline ¡n the prevalence in follicutar trachoma was found. An odds ratio for years lggg/go was oR = 2.0, 95% C.l. 1.64 to 2.41. The differenCeS between the are

presented graphically in Figure 4'3.1 .

189 Chapter 4

Figure 4.3.1: lnflammatory lrachoma 1976 compared to 1989/90

80

-o- 1 I 76 60 + I 989/90

% of age-grouP with ¡nflammatory 40

trachoma

20

0 ot ?3 +s ô7 80 '1011 12'13 l+15 ls17 l&19

Age-group

190 Chapter 4

prevatence rates ol Blindness lor the (c\ Prevatence of Btindness: A Comparison between the

R.A.C.O.Reportandthel989/90surueyl4edCentre'zonedata

for The prevalence rates for btindness were compared using R.A.C.O. Report 1980 definition the age /sex blindness (visual acuity in the better eye of 6/60 or worse. Table 4.3.8. outlines

distribution of blind individuals in the "Red Centre" zone'

survey Table 4.3.g.: Number of Blind lndividuals for the "Red Centre' zone using 1989/90

data (N.T.E.H.P. defn.)

Age F M Total

0-19 0 0 0/610

20-29 0 1t'177

30-59 0 0 0t312

60+ 16 5 21t120

17 5 22t1219

Four comparisons were made of the blindness prevalence rates using the 1980

R.A.C.O. Report and the 1989/90 survey data. Crude proport¡ons were used in testing the

significance of the changes. Prevalence rates per thousand are also presented (Table 4.3.9).

191 Chapter 4

Table 4.3.9.: Prevalence ol blindness: A comparison between the 1980 R'A'C'O

RePort and the 1989/90 SurveY

Prevalence Rates Type of ComParison P 1 980 1989/90 f

1.09 0.29 Crude prevalence from 183t7997 22t1219 lrom survey data Rates per 1000 23 18 0.03 Prevalence using the source 23t1265 22t2386 4.76 population as a denominator Rates per 1000 18 10

Prevalence in the 60+ age 157t692 21t120 1.32 0.25 using 1976 census data Rates per 1000 227 172

Prevalence in the 60+ age 20t78 21t147 3.77 0.05 using the source PoPulation as a denominator Rates per 1000 256 143

The results indicate a reduclion in the prevalence of blindness in the populat¡on as a whole

(using the source population as a denominator), if assumptions about the unsampled

population were correcÍ.

When just the 60+ age-group was considered, male/female data was available from the

R.A.C.O. Report (1980 p62). ln the 'Red Centre" zone 58/313 (185/1000) males were

binocutarty btind compared to 98/379 (259/1OOO) females. A ditference that was signilicant

(p=o.OZS). These rates can be compared to the 1989/90 data: 5/55 (91/1000) and 17165

(26i11OOO) respectivety (using the source population as a denominator 5n0 V111000) and for males fln5 eßt1ooo)). A significant reducfion in the prevalence of blindness oocurred source us¡ng the source population (p=0.02), but not for females with either the sample or

denominators.

192 ChaPter 4

(3) 1985 N.T.E.H.P. Revlew Data (T.E.H.P. Report)

(a) comparison of the poputations for the 1985 N.T.E.H.P. review and the 1989/90 suMey.

Table 4.4.0 compares the source population for the communities surveyed ¡n the 1985 19gS N.T.E.H.p. review with the same communities in each of the years 1976,

and 1999/90. As estimated from 1976 census data, 1986 census data, and 1989/90

poPulation lists.

The communities surveyed in the 1985 N.T.E.H.P. review were Yalata, lndulkana,

Amata and Fregon.

1986 Tabte 4.4.0.: Comparison of the source population : N.T.E.H.P. review 1985 (1976 and

census data) w¡th the 1989/90 survey (1989/190 populalion list data).

1989/90 Age-group 1 976 lo ol 1 986 lo ol (Years) Census Total Census Total Population List

0-4 153 13.8 129 11.4 201 12.1 5-9 145 13.1 109 9.6 221 13.3 10-19 230 20.1 277 24.4 376 22.7 20-29 181 16.4 228 20.1 300 18.1 30-39 180 16.3 137 12.1 201 12.1 40-49 99 8.9 90 7.9 131 7.9 50-59 62 5.6 73 6.4 113 6.8 60+ 56 5.1 93 8.2 117 7.0

Total 1106 100o/o 1136 1000,6 1660 1000/6

Note: ln f 976 the communities of Amata, Fregon and Pipalyatiara were part of one collector

district. Adjustment was necessary using 1981 census data lo obtain an approximate totalfor

the communities of Amata and Fregon alone.

193 Chapter 4

were very comparing the population structures of the source population it appears they similar

survey Table 4.4.1.: Comparison of the samples : N.T.E.H.P. Review 1985 with the 1989/90

o/o 1989/90 ol Agegroup 1976 1" ol 1 985 ol "h Total (Years) Sample Total Sample Total Sample

110 13.0 0-4 94 12.9 28 11.4 15.9 s-9 1s1 20.7 s9 24.1 136 94 1 1.1 10-14 101 13.8 82 33.s 75 8.9 15-19 65 8.9 14 5.7

0-19 411 56.2 183 7 4.7 415 48.6

25.2 20-39 162 22.2 24 9.8 21s 15.9 40-59 s9 12.2 19 7.8 136 10.3 6Où 69 9.4 19 7.8 88

Total 731 I 00% 245 I 00% 854 I 00%

ln The sample totals for the 1976 and 1990 surveys were s¡m¡lar: 731 and 854 respectively.

the 19g5 survey, only 245 individuals were examined. The maþrity were in the 0-19 age

bracket. The age structures of the 1976 and 1989/90 surveys were comparable, but the

1985 sample ditfered from both the 1976 and 1989/90 surveys'

(b) comparisons Based on the 1985 Trachome and Eye Health Program Review

(N.T. E.H. P. Revi ew) : follicu I ar t ncho ma

The prevalence of follicular trachoma for the commun¡ties of Yalata, lndulkána, Amata and

Fregon for 1976, 1985 and 1989190 were compared.

194 Chapter 4

4.4.6. for the Tables 4.4.2. - 4.4.5. give the results for individual communities and Table for the 1985 communities combined. The delinition of lollicular lrachoma was standardized The report (section 4.2.), the results for 1976 and 1985 are therefore directly comparable. would overlap with defin¡t¡on of foilicutar trachoma (TF) used in 1985 included a category that

the current grading of trachoma intense (section 4.2). The results for the 1989/90 survey and were for total active inflammatory trachoma (trachoma follicles and trachoma intense), result included an additional category of trachoma grading for single eyes' The f 989/90

would therefore overestimate the prevalence of trachoma compared with the 1976 results.

Table 4.4.2.: Folliculartrachoma in Yalata by year and age

1989/90 1 976 1 98s

oy'o Total lo Age- Foll. Total fo Foll. Total fFm Group Trach. Examined Trach ExaminEd ExaminEd

32 15.6 0.4 10 24 41.6 1 2 50 5 52 19.2 5-9 18 43 41.9 13 32 40.6 10 45 8.8 10-14 I 28 28.6 4 27 14.8 4 27 0 1s-19 3 22 13.6 0 1 0 0

o-19 39 117 33.3 18 62 29 19 1s6 12.2

9.5 20-39 4l 2.5 0 1 0 7 74 8.2 40-59 0 24 0 0 0 4 49 18.8 60+ 0 23 0 0 5 0 3 16

Total 40 207 19.3 18 69 26-1 33 295 11'2

yalata showed an overall decline in the crude prevalerrce of TF lrom 19.3% in 1976 lo 11'P/o

in 1990. The difference was statistically significant. The age-standardized odds ratio of

follicular trachoma when the results lor 1976 were compared to 1989/90 was: M-H weighted

195 Chapter 4

¡n the o-19 oR=1.86, 95% C.l.: 1.08 to g.27, (xz = 5'05 (p = 0'025)), us¡ng the age-groups The increase in 1985was age bracket and the aggregate resutts lorthe 20-60+ age-group' comparable with the based on a small sample of young people, and was therefore not directly other years

from 33'3% in Age specific rates showed a statistically significant decline in 0-19 age-group increase in inflammatory 1976to 12.2ohin 1990 1X2 = t6.64 (p < o.oo1)). conversely an (p=O'016))' This may trachoma (TF + Tl) occurred in the 20+ age-group in 1989/90 1X2= 5'ZB definition, or to have been artelactual, due either to the inclusion of trachoma intense in the

grading TF when TS was possible. Alternatively it may rellect a true change in the age where results distribution of the d¡sease. This was supported by data from the interim reports in for follicular trachoma (without the addition of Tl) showed an increase ¡n the 20+ agegroup

the 1990 survey compared with the 1976 survey'

196 Chapter 4

Table 4.4.3.: Follicular trachoma in lndulkana

1 1 976 1 985 989/90

o/o o/o Total Age- Foll. Total Foll. Total % TF/TI Group Trach Examined Trach Examined Examined

04 7 12 s8.3 6 o 100 4 27 14.8 '15 27 37.0 5-9 31 42 73.9 7 46.6 10 21 23.8 10-14 18 35 51.4 2 13 15.4 5 0 1s-19 3 5 60.0 0 3 0 0 18

0-19 59 94 62.7 15 37 40.5 19 9 3 20.4

20-39 42 2.4 0 0 3 4 6.8

40-59 2 22 9.1 0 3 0 5 26 19.2 15.8 60+ 0 11 0 0 0 0 3 19

Total 62 169 36.7 1 5 41 36.6 30 182 16'5

lndulkana had a high prevalence of follicular trachoma in 1976, particularly ¡n the 5-9 age'

group. This had lallen by 1985, although only age-specific rates were comparable.

The prevalence of follicular trachoma also had fallen in 1989/90. The ditference was

statistically signif¡cant. The age-standard¡zed odds ratio ol follicular trachoma when the results

for 1976 were compared to 1989/90 was: M-H weighted OR=2.55, 95% C.l': 1.54 to 5.25, (X2

= 11.92 (p < 0.001)), us¡ng the age-groups ¡n the 0-19 age bracket and ths aggregate results

for the 20-60+ age-group. But again an ¡ncrease in the crude follicular trachoma preva¡ence

was seen in the 20+ age{roup, but this was not significant (X2 = 2.65 (p=0.100)).

197 Chapter +

Table 4.4.4.: Follicular trachoma in Amata

1 989/90 1 976 1 985

TFM Total o/o Age' Foll. Total fo Foll. Total % Examined Group Trach. Examined Trach. Examined

15 28 s3.6 04 18 27 66.6 6 7 85.7 29 31.0 5-9 24 37 64.8 7 I 77.7 I 42.8 2 22.2 10-14 7 18 38.8 6 14 I 0 't6 0 5-19 2 26 7.7 2 6 33.3

26 82 31.7 0-1 I 51 108 47.2 21 36 58.3

48 4.2 20-39 0 41 0 0 11 0 2

1 26 3.8 40-59 0 22 0 0 5 0

7 0 1 2g 3.6 60+ 1 13 7.7 0

Total 52 184 28.3 21 59 35.6 30 184 16.3

The results lor Amata, like lndulkana, showed a decl¡ne ¡n the crude foll¡cular trachoma

prevalence The difference was stat¡st¡cally signiÍicant. The age-standardized odds ratio ol weighted foilicular frachoma when the results for 1976 were compared to 1989/90 was: M-H age oR=2.1g, 95% C.t.: 1.12 ro 4.33, (X2 = 5.24 (p = O.O2)), us¡ng the age-groups in the o-19 Fregon and bracket and the aggregate results for the 20-60+ age'group. But, Amata unlike 20+ age- lndulkana did not sho¡r a large increase in crude follicular trachoma prevalence in the groupandthedifferencewasnotsignificantlX2=1.08(p=g.2gFisherexacttest)).llthe variation increase was solely due to systematic grading error, or differences in definition, the pattem would be expected to be similar across all communities. Therefore a changing of

disease may explain at least some of this increase.

198 Chapter 4

Table 4.4.5.: Follicular trachoma in Fregon

1 989/90 1 976 1 98s

Total ot Age- Foll. Total Foll. Total % TFÆI Group Trach. Examined Trach. Examined Examined

23 39.'l 0-4 15 31 48.4 10 13 76.9 I 5 28 17.9 5-9 11 29 38.0 0 3 0 19 10.5 10-14 7 20 35.0 11 28 39.2 2 0 14 7.1 15-19 1 12 8.3 0 4

0-19 34 92 37.0 21 49 43.7 17 8 4 20.2

49 4.1 20-39 0 36 0 0 11 0 2 s.7 40-59 0 21 0 0 10 0 2 35 25 8.0 60+ 0 22 0 0 7 0 2

Total 34 171 19.8 21 76 27.6 23 193 11-9

Fregon had almost identical overall results to those of Yalata, but, the observed decline in

crude follicular trachoma from 34/171 lo 231193 was not stat¡stically significant. The age-

standardized odds ratio of follicular trachoma when the results for 1976 were compared to

19g9/90 was: M-H weighted oR=1.15, 95% c.l.: 0.77 lo 2.gg, (x2 = 1.25 (p = o'2æ)), using

the age-groups in the 0-19 age bracket and the aggregate results forthe 20-60+ age-group. (X2 (p For the 0-19 group the decl¡ne from 34/92 lo 17lL4was statistically significant = 5.18 =

0.023)). As with Amata, the increase in the older age-group was less than that occurring in

Yalata and was not significant (x2 = 1'27 (p=o'lg Fisher exact test)) The persisting high years prevalence in the 0-4 age-group was a cause for concern. The ditference between the

in Yalata was not significant (X2= 0.45 (p=0.502)). A similar picture occurred in Amata but not

or lndulkana.

199 Chapter 4

Table 4.4.6.: Follicular trachoma in all the communities combined

1 989/90 1 976 1 985

o/o TF/TI Total fo Age Foll. Total lo Foll. Total Examined Group Trach. Examined Trach. Examined

110 30.0 0-4 50 94 s3.2 23 28 82.1 33 34 136 25.0 5-9 u 1s1 55.6 27 59 4s.8 94 13.8 10-14 40 101 39.6 23 82 28.O 13

1 7S 1.3 15-19 9 65 13.8 2 14 14.3

0.1918341144.57518341.08141519.5

2',15 6.s 20-39 2 162 1.2 0 24 0 14 8.8 40-59 2 89 2.2 0 19 0 12 136 19 0 9 88 10.2 60+ 1 69 1.5 0

I 13'6 Total 1 88 751 25-7 7 5 245 30'6 16 854

The combined results averaged the observed vadation between commun¡t¡es' There was a

statistically signilicant decrease in the crude trachoma prevalence in the 0-19 age-group

20+ IS.SA

lollicular trachoma when the results for 1976 were compared to 1989/90 was: M-H weightsd in Table 4'4'6' Q+=2.24,g5o/o c.l.:l.62 to 2.90, (X2= 28.76 (p <0.001)),using the age-groups

may As has already been discussed for the individual commun¡ties, the inclusion of Tl with TF

explain some of this increase. However, a change ¡n the pattern and d¡str¡bution of years inflammatory disease cannot be exctuded as a cause. The ditferences between the is

denþnstrated in figure 4.3.2 .

200 Chapter 4

Figure 4.3.2: Follicular trachoma by year (1976,1985 and1989/90)

100

80 + I 976 + 1 985 + 1 989/190 % of age-grouP 60

with follicular 40 trachoma

20

0 04 59 l0l4 lS19 20æ 4o5g S+

Age-group

201 Chapter 4

for the three years using togistic regression analysis a model was derived of the results

Follicular trachoma was found to vary by group and year'

For this model we have the following est¡mates'

Estimate SE t statistic

-18.409 Constant -3.383102 0.183777 13.059 Group 1 2.741640 0.209946 Group 2 2.5f 8410 0.196422 12.821 Group 3 1.368844 0.2032537 6.7347 Group 4 0.0 0.0 0.0 5.0035 1 976 0.713983 0.1426969 4.8694 1 985 0.9155231 0.1880173 1989/90 0.0 0.0 0.0

Where group was representative of 4 age categories l\2,7, 15,451'

A reduction in the prevalence of follicular trachoma was demonstrated. An odds ratios

between the years 1976-1989/90 was: oR = 2.04, 95% C.l.: 1.54 to 2'68.

202 ChaPter 4

Summary

Pitjantjatiara' Four sets ol data were used to review the changes in eye health in the Anangu for the years from 1976 to 1989/90.

(1) The interim rePort data.

(2)The.RedCentre"zonedatafromtheR.A.C.O.Reportofl9S0.

(3) The data from the N'T'E.H.P. review 1985'

(4) The 1989190 surveY data.

Trachoma young' The prevalence of follicular trachoma was found to have decreased significantly in lhe

in the years from 1g76 to 1989/90, for both the Crude or Overall rates and forthe age-

standardized rates. However, there was an increase in the prevalence of inflammatory

trachoma in the older age-groups.

(1) lnterim reports : A reduction in the crude and age-standardized folliq.llar trachoma

prevalence rates from 1976 to 1989^90.

: An increase in the prevalence of folliqJlar trachoma in the 20+ age'

group from 1976 to 1989/90.

(21T.E.H.P. Report : A decrease in the age-specific prevalerrce of follicular trachoma as

measured by, TF in 1976 and TF+TIin 1989/90. Age-specific rates

were shown to have decreased in all age-groups except the 0-1, 10-

11 and 18-19 age-groups.

: Logistic regression analysis confirmed the observed decline in

follicular trachoma for the 0-20 age-group in the years frcm 1976 to

1989/90.

(31 1985 Review : A reduction ¡n the age-standardized prevalence of follicular

trachoma as measured by, TF in 1976 and TF+TIin 1989/90, was

203 Chapter 4

found in the communities of Yalata, Amata and lndulkana, but not in

Fregon.

: An increase in the crude follicular trachoma prevalence for the 20+

age-group, in the community of Yalata, but not lor the other

communit¡es.

: A decrease in the age-standardized prevalence rate of folliqjlar

trachoma for the combined communit¡es. Ho¡rever, an increase was

found in the 20+ age-group.

VisualAcuity T.E.H.P. Visual acuity was compared using data available from the interim reports and the report (1985). poor vision was (11 lnterim repofts : A decrease in the age-standardized prevalence of

found in resuJts for the communities combined.

: A decrease in allthe age-specific prevalence rates of poor vision,

except in the 61+ age-group.

: No reduction in the crude prevalence of blindness for the combined

communities

population prevalence of blindness l2''l R.A.C.O. RePort : A reduction in the source

(N.T.E.H.P. defn.), but no significant decrease found for the sample'

: A reduction in the prevalence of blindness in males but not females

using the source population as a denominator.

: A signilicant difference ¡n the proportion of males and females

binocularly blind. which was consistent with the results ol the

1976-79 N.T.E.H.P. survey.

204 Chapter 4

4.4. Dlscusslon

Although there were methodological problems when the 1976 N.T.E.H'P. survey was from the compared to the lggg/go survey, there was consistency in the results obtained lhree comParisons used.

The two principal methodological problems were the different grading systems and the be made, selection ol the sample. Although the grading systems differed comparisons could relative to acknowledging that the 1gg0 criteria of ths 5 sign system would overest¡mate TF

purposes Tl + the 1gg5 and 1g80 definitions (section 4.2.), especially when for comparative

TF were combined to generate rates for 1989/90 data'

The sample and source population of the surveys were compared and found to be

approximately equivalent (section 4.3.). This could be expected because although lhe

population had grown there had been no documented net migration. The communities were participation basicalry the same, except for the growth of homeland settlements and the by

people in the survey would be expected to bo the same because the need for eye care was

unchanged. Additionally the purpose and extent of each survey was similar.

The following hypothesis was tested

"The age-specific prevalence rate of blindness and trachoma in the Anangu Pitjantjatjara of

South Austral¡a have shown a statistically significant decline since the N'T.E.H.P. survey of

1 976".

The results of the comparisons (section 4.3) between 1989/90 and 1976 indicate that lor

follicular trachoma, a stat¡stically significant reduction was found lor all age-standardized

prevalence data from the interim reports, from the "Red Centre" data for the 0-20 agegroup'

and from the t9B5 Trachoma and Eye Health report for allthe individualcommunities (excefl

Fregon) . The declÍne was also evident from the comparisons using logistic regression

analysis. However, it was noted that the age-specific prevalence rates for inflammatory

205 Chapter 4

previously this may be due trachoma in the older age-groups (20+) increased. As mentioned surveys (e'g' to the inclusion of Tl with TF, or the differences in grading TF between the survey but not individuals with TF in one eye only were counted as having TF in the 1989/90 a change in the in the N.T.E.H.P. survey or the 1985 review). Atternatively it may represent trachoma in the distribution of disease, such that, a cohort of individuals severely atfec{ed by 1990s, wh¡lst a 1970s have grown older and continue to sutfer from poor eye health in the younger generation grows up with less severe disease. A reduction in inflammatory trachoma A review found in the younger age-groups has been found before in the 'Red Certre" zone' in the a statistically significant reduction in TF lrom 1976 to 1985 for the 0-9 age-group

Warburton Range of Western Australia (Cooper et al. 1986). The authors thought the 1985 N.T.E.H'P' avaitability of water in the region may have contributed to the decline. The lndulkana' review found no statistically significant ditference in the prevalence of trachoma in was an Fregon and Amata when 1976 was compared to 1985 screenings (although there Abodginal apparent rise in Amata and Fregon with a reduction in lndulkana). Other studies ol decline in communities in centralAustralia by Meredith et al. (1989), have not demonstrated a

active i nf lammato ry trachoma.

performed, as was Recalculation of the trachoma rates for the source population were not infection done for poor vision and blindness, because there was no evidence that the rate of

in the unsampled populalion was any ditferent from the sampled group.

data' Unfortunately, due to problems of disease definition and access to original age-specilic

prevatence rates for trachoma scarring, trachoma trichiasis and corneal opacity were not "Red compared. Comparison of the crude cicatriciat trachoma rates (TS + TT + CO) for the

centre" data gave a guide to the trend. For all ages the rate were 661/1017 in 1976 versus

3g4t1218in 1989Æ0, a statistically signilicant reduction 1f = StS.S (p < 0.OO))'

section From the qualitative examination of past surveys in South Australia (section 2.4 and during 2.5), ¡t appears that there was an initial rise in the prevalence and severity of trachoma

the 1960s and 1970s, foilowed by adecline inthe 1980s. Such a rise andthan alall inthe

206 Chapter 4

trachoma prevalence and severity of trachoma could be consistent w¡th a late introduction ol

infective disease in a into these N.w. communities, with trachoma initially causing a high rate of detectable. community not previously exposed, but with little secondary severe disease untreated Then over time some individuals develop the blinding complications ol chronic process. A disease. The surveys in the late 1970s may have detected this phase of the programs gradual improvement in economic development, living conditions and treatment pool atfected produces a decline in the active infection rate, but w¡th a residual of severely 1980s' people, who form a cohort, and are detected by each subsequent survey in the

high rates ol A similar but reverse pattern of disease occurred in southern Malawi, where was lrachomatous inflammation in children and high rates of cicatrizat¡on in adults the authors accompanied by the absence of trachomatous blindness. lt was hypothesized by had that in the valley studied, dramatic changes in the local ecology towards desertification

produced a cohort etfect; a group of older people, previously not aflecited by trachoma, now group to have existing in an environment favouring infection with trachoma. This had survived progress to rnore the first changes towards blindness (cicatrization), but had not had time to years severe complications. lf high rates of inlection had been present for 30 or more a

greater prevalence of blindness due to trachoma would have been lound (Telsch et al. 1988).

The comparisons between 19gg/go and 1976 demonstrated a statistically significant decl¡ne

in poor vision and blindness. Two assumptions were made in dedving prevalence rates for

comparison. The first assumplion was the extrapolation of the "Red Centre" data from the

R.A.C.O. Report of 1980, to a sub-set of data for the Anangu Pitjantiatjara lands. The second population' assumption was that there were no additional blind individuals in the unexamined

Support for both these assumptions was lound in:

(a) The rate for binocular blindness recorded for the interim reports that were

(a prevalerrce avaitabte for the 1976 survey. Blindness was found in 19 of 698 people no rate of 27|1OOO forthe sample or 18/1000 forthe source population, assuming

additional blind individuals). This may be compared with the rates for the 'Red

207 Chapter 4

these Centre" zone data ot 23t1000 and 18/1000 respectively. The slmilarity of to figures supports the assumption that the "Red centre' zone data could be used

lands. form a set of data for comparison with the data for the Anangu Pitjantiatiara

(b) The assumption that all or nrost of the binocularly blind individuals were seen

was based on the high sampling in the aþrisk group aged 60+ (79"/"1, examination of population medical records for individuals not seen in the survey but identified by the

lists, and the use of local health workers to identify blind individuals'

known similar assumptions could be made about the 1976-1979 survey; however, it waS not data, the whether a thorough search was made for blind individuals. But, using 1976 census interim reports demonstrated an overall participation of 66% (668/1052), with a 100% participationl in the 60+ age-group, the group with the greatest proport¡on of blindness.

Although it could be speculated that most of the blind choose to be examined, and the cooperation of the communities in enabling this to happen was effective, Prof. Hollows did

comment'the blind were otten immobile wh¡le those with visual impairment were often unable

or unwilling to attend, whether because of ditfidence or shame, or because they had (R'A.C'O. accepted their disability and were pessimistic about the prospec'ts of improvement'

Report 1980 P13).

lf allblind individuals were examined in the 1976-1979 survey any prevalence figures based

on the sample would be an overestimate of the true prevalence, only by using the source

population as a denominator would the estimate be accurate. Conversely, if many blind

individuals were not examined the estimate using the sample could underestimate tho true

prevalence of blindness, and using the source population would reduce the estimate lurther.

Comparing the prevalence of blindness using the above assumptions dernonstrated that the

reduction in the prevalence of blindness, from 1976 to 1989/90, was statistically significant

1 More people were examined than seen in the 1976 census

208 Chapter 4

(section 4'3')' using the source population as a denominator, but not for the sample of the true However, the source population estimate for 1989/90 may be an overestimate

population. lf the estimate for the source population was reduced by 10%, (the figure derived

but were still significant (1or the f rom the survey in Ernabella (seaion 3.3.1) the results change 1986 census data source poputation 29¡265 vs 22t2147, (X2=3.85 (pc0.05))). lf proiected 'Red was used the result was not significant (p<0.05 level). When the 60+ age-group for the

had occurred for males Centre" zone was considered (section 4.9. (2)), a significant reduction

but not females.

The difference lound in the source population prevalence rates lor blindness was supported (section crude by the data lor poor vision. When the two surveys were compared 4.3.) the

prevalence of poor vision for the sample showed a stat¡st¡cally significant ditference between

the years, the age-standardized odds ratio between the years being oR=2'86, 95% C.l.: 1.86

to 4.75, (X2= 23.2 (p<0.001)), us¡ng the age-groups in the 21-61+ age bracket and the

aggregated results lor the 0-20 age-group. Statistically significant differences were also

found for all bar the 60+ age-group for the sample, or for the 50+ group using the source

population as a denominator. The 60+ age-group was aged 45+ in 1976, and probably had a

high rate of visual impairment then, although individuals who were blind or severely visually

impaired in 1976 may have had higher mortality rates than those with good vision, a cohort

effect could be maintaining the prevalence rates lor this group at a high level.

The overail reduclion demonstrated across all age-groups was not due to an increasing

proportion of young people (who have good visual health). ln fact the proportion of "aged"

(50+) to "young, (O-49) had increased from 1976 to 1989190, (10.8% lo 13.7'h (source) and

'11.6%to 1 6.6% (sample) respectively).

The cause for the decline in visual disability may be related to the decline in inflammatory

trachoma, w1h fewer young individuals progressing to poor vision and blindness because of

repeated infection. This was demonstrated in the table showing individuals at risk of binocular

blindness from trachoma (Table 3.3.5.3.). Ol 15 individuals at high dsk of blindness from

209 Chapter 4

individuals with binocular trachoma out of 66, only 1 was aged less than 60. of those one was aged less than 60. blindness of whom nine were blind due to trichiasis, again, only percentage of the young Additionally, trichiasis and corneal opacity were found in only a small (\0'79%) and 5/1134 (less than 40). For ages less than 50 the proportion were 9/1134 trachoma (0.44"/.1respectively. lndicating that few individuals were at risk of blindness due to

in the < 50 age-grouPs'

l¡festyles which The decline may also have been due to better living conditions and changed of monocular might decrease lhe risk of traumatic visual loss. Trauma accounted lor 28'8lo probably small (Table blindness but its contribution overall to binocular blindness was 3.3.4.3.). Another factor which could have a maior impact on the prevalence ol visual areas of tertiary ¡mpairment, would be the availability of surgery for cataract and trichias¡s, two However, it treatment which had been provided more consistently since the 1976 survey' of binoqrlar was alarming that cataract accounted for 97.g%of monocular blindness and 40!" visually blindness. ll the survey were repeated in 1991 many ol the 33 individuals severely operated on, impaired because of cataract diagnosed in the f 989/90 survey would have been prevalence rates, decreasing ánd would no longer contribute to the poor vision or blindness medical recods the rates further. A review of the cataract operations and their oulcome, using principal for tertiary at the Royat Adetaide Hospital and Ftinders Medical centre (the centres health' treatment), would provide uselul incidence data on these aspects of Anangu eye

factors; Why has this decline occurred? The survey was not designed to identify specific faclors, however, several explanations, including cultural, environmental and socioeconomic areas of can be advanced (section 1.3 (2.1.7 (b) and (c) )). Declines have been found in (42'399 previously high endemic1y. One survey of hospital outpatients statistics in Nigeria morbidity of cases) demonstrated a prevalence of trachomatous blindness of 0.5% and ocular years ago (Bababla 0.5% in 1g86 compared to figure s ol 21-44o/o and 23o/o respectively thirty

19gg). The author acknowledged the shortcomings of an outpat¡ent survey but suggested topical the declfne in lhe importance ol trachoma may have been due to the availability of

210 Chapter 4

New Guinea antibiotic medication and improved living standards. Another survey in Papua which was compared to the results of Mann in 1955, lound that atthough trachoma was stil! endemic, it was reduced in prevalence and severity. lt was thought improvement in standards of personal care and community hygiene contributed to the decline (Dethlefs 1982). ln a similar fashion, community and socioeconomic development on the Anangu Pitjantjatjara lands may have contributed to the decline. Events such as the granting of the free-hold title of the lands to the Anangu Pitjantjatjara, the establishment of a self-governing body

(pitjantiatjara council lnc.) and the organization of an independent health service (Nganampa

Health Council) may have helped. An increase in the quality of housing has occurred, and there has been a trend to the settlement of homeland areas, which promote traditional

Anangu ways, and a decentralization of the population. Employment opportunities have increased (although still small by comparison with large cities) and educationalopportunities

have also broadened. The availability and access to water has improved, with more homes

having reticulated supplies. This has probably been combined with a change in personal

hygiene practices, epitomized by recent advertisements aimed at children on the local Central

Australian Aboriginat tetevision (C.A.A.M.A.), of the fun and importance of showering and

washing faces. The world literature supporting the role of increased access to water in

decreasing the prevalence of trachoma, with hygiene lactors also being ¡mportant (section 1.3

(2.1.7 (e) )). Access to medicalfacilities has also improved with the opening of clinics in allthe

major setlemenls, which can also provide referrals for tertiary treatment at Alice Springs or

Adelaide. Of importance to eye care specifically has been the South Australían Aboriginal

Trachoma Program visits which have identified and initiated trachoma treatment, partiollarly in

children

211 Chapter 5

Chapter 5. Research lssues

212 Chapter 5

5.1 lntroductlon

The preceding chapters have detailed the formulation of a population based cross-sectional prevalence survey in the Anangu Pitiantiatjara, and the results of comparisons with previous surveys.

There were three principal research issues' These were:

1. The issues associated with the 1989/90 survey. Firstly, the practical problems

encountered in pelorming the survey were not unique to research in Aboriginalcommunities

and have ramifications for research in any defined communily. Specifically, it requires

researchers to be aware of cullural ditferences and to allow for them in study design.

Secondly the 1g8g/go survey analysed observer error in a manner different f rom previous eye

surveys. Thirdly, the etfect of non-magnification on trachoma grading was assessed because

loupes were occasionally ditficutt to use in lhe lield and it was some health workers were noted

not to use them when assessing trachoma.

2. The methodological problems ol the survey and the comparison was another issue

which was important to the validity of the conclusions.

3. The issue of medical and public health research in Aboñginalcommun¡t¡es was both

topical and evolving during the course of the study. Additionally, there were specific ethical

considerations associated with the performance of the survey and the d¡stribut¡on of the

results.

213 Chapter 5

5.2 1989/90 SurveY lssues

5.2.1. Practlcal Problems ln the Communltles

During the survey several pract¡cal problems arose, which although they atfected the conduct

of the survey, did not significantly affect the results. The problems reflected the cultural priorities differences between the survey team and the Anangu P¡tiant¡atjara, and the of the

local commun¡ties. At alltimes both groups were pat¡ent and cooperative enabling the urvey

to proceed with minimaldisruption.

Deaths in the Commun¡ties

On at least two trips to communities, a recent death adversely atfec{ed part¡cipation in the

survey . A death woutd often involve the whole community and mean the movement of

people away from either the house or area where someone died, with perhaps the *sorry establishment of camps" where lhe process of mourning would occur. An "expected-

death of either avery old person or sick individual usually resulted in less disruption to the

community than the sudden and tragic loss of a child or formerly healthy adult. However, a

unique combination of factors occurred on each occasion, and although some generalisations

coutd be made about outcomes, these were not stdctly predictable. For example, although a

community might be in nnurning, partic¡pat¡on ¡n the survey might could stilloccur.

Football

Australian Rules Footbaltwas played with enthusiasm by many young men. Each community

had a team and games were played on earthen playing lields during the w¡nter months.

Matches were played between communities within the Anangu Pitiantiatjara lands, but could

also involve Aboriginal people from other CentralAustralian areas. The activities were often

broadened to include other games such as basketball.

214 Chapter 5

significant Families would accompany the playing members of a team and the occasion was a game' social event. The population of a community could double or treble if it was a weekend population in Conversely, the other communities could lose a significant proportion of their

the days leading up to a carnival and those lollowing it. Therefore, the t¡m¡ng of a survey team

visit to a particular community was important, avoiding if possible an away match, or at least in staying for a longer period. Unfortunately the timetable for matches was ditfict¡lt to ascerlain

advance, and could change at short notice. Therelore, it was necessary to stay for as long as

possible and have a flexible attitude to the inevitable changes in timetable.

Ceremonies

Throughout the calendar year but perhaps more towards October and November, important

ceremonies were conducned throughout the Anangu Pitjantjatjara/Yalata lands. These

ceremonies were centralto the functioning of the community and had precedence over other

activities. The conduct of the survey at these times was difficult but not impossible. The

communities and individuals were very co-operative and helpful when the survey visits

coincided with initiation or other ceremonies. During interludes from the ceremonies, when

people returned to the central community, individuals were otten happy to be examined; and

on one occasion the survey team was invited to visit a camp of several hundred people to

examine anyone who wíshed to be tested.

Again, atthough ¡t ¡s easy to generalise, each occas¡on was different and the team was guided

by discussion with health service statf, the community council, and the elders who lived in the

commun¡ty. Consuttation was the key element in obtaining permission to cont¡nue the survey

at these times.

Absence ol Young Men

An observation concerning the relative absence of young men was corlirmed ¡n the f¡nal

statist¡cs. Although it was evident from the populat¡on l¡sts that about as many young men

aged 20:30 lived in the communities as did young women, their participation ¡n the survey was

215 Chapter 5

good health and limited. several reasons can be proposed for lhis. Most young men had eye probably d¡d not feel the need to have their eyes tested. The young men were also otten or involved in football or ceremonies. There may have been some element ol embarrassment shame in having their eyes tested or being compared to others. Lastly, anecdotal evidence suggested that they tended to move around the communities more frequently than other age-groups. They were also away hunting or visiting regional centres such as Uluru, Alice

Springs or port Augusta, and therefore not be available to be approached lor testing.

Although it was disappointing not to sample this group comprehensively, because of their generally good eye health they would have contributed little to the morbidity statistics for trachoma or blindness, and by using rates based on the source population the contribution of this age-group was taken into account.

Nomadic Lifestyle

prior to white setflement, the Anangu Pitjant¡atjara moved across the landscape visiting krpwn

water-holes, hunting areas and cerernonial sites according to the time of year. Allhough nnst

people now identify with a particular community, time is otten spent at other comrnun¡ties or on

the homelands with extended family members. These periods away could last several

months, therefore it was not uncommon to find several individuals visiting or spending

extended periods in other communities during survey visits.

It was important, therefore, to establish which community individuals identified with, ¡denl¡fy

them on the poputation list and avoid duplication. Double counting, part¡cularly of those with

eye disease, theoretically could have led to significant inaccuracies in the data.

Weather

ln general the weather was mild; however, both extreme heat and cold had negative eflects

on partícipation in the survey. On one survey trip, the cold and harsh weather reduced the

number of people visiting the clinic. Because testing occurred outside parlicipation was

further curtailed and the examination made more ditficult because ol the windy conditions.

216 Chapter 5

well equipped clinic Rain on another occasion had a similar eflect, despite having a large and to leave their house and being able to conduct the exam¡nation inside . People were unwilling etfect and or camps and brave the elements just for an eye survey. Extreme heat had a similar sapped the energy and enthusiasm of the survey team'

Cultural

team on There was an obvious gap between the culture, expectations and aims of the survey the one hand and the Aboriginal people on the other. Although the survey had a strong people possible' The service element, the underlying aim was to see and examine as many as people survey and its role was discussed with the community council and explained to of the attending for examination. The involvement ol the councilwas the key to lhe success single survey in any community. Also it was found that a multiplier effect occurred when a family member was examined and the importance of the survey explained to him or her. Otten s/he would lhen encourage other family members to attend for examination. lt was the establishment of a relationship with the first person which facilitated the participation of others.

The importance of relationships to Aboriginal people recurred again and again, with

grandparents for whom spectacles had been prescribed encouraging younger family people members to attend, or with brothers and sisters bringing siblings. Sometimes would

attend not out of any perceived setf-need but because they thought the survey was important

lor the community. On other occasions they did so because they wanted to help the survey

team as peoPle not as researchers.

Eventually, after several trips some of the survey team became well known and friendships

developed on previous visits helped participation on subsequent trips.

Resources

Although the survey team was effectively self-contained ¡n terms of equipment, it was reliant

on a clinic or other space lor the examination. The quality of lhe accommodation vañed

enormously from spacious air conditioned clinics, to open areas under trees. lt would have

217 Chapter 5 been useful to be completely self-contained to avoid disrupting the clinic routines; however, the resources for the survey could nol be extended to meet such requiremenls..

Timetables

Timetables were perhaps the greatest source of frustration in organising the survey. Ol first priority was the communities' preferences, each of which had a number of impoilant prograrns occurring throughout the year, along with events such as football, business and ceremonies.

The dates given then had to be coordinated with the Trachoma Program, the Royal Flying

Doctor Service (R.F.D.S.), who provided lransport, and the ophthalmologists' Occasionally

trips had to be cancelled or rearranged. However, as long as sufficient time was eventually

spent in each community problems did not arise. The best part¡c¡pation rates occurred in

communities which either had extended visits, such as Ernabella (1990), or communities

which were visited on at least two occasions. ln each case the survey team became familiar

wilh the physical surroundings, health personnel and the people. Similarly, the people had

time to familiarize themselves with the survey team and had time to make their own way to the

clinic for review. On such trips, repeated visits to the schools often increased the numbers in

the younger age-groups significanlly, because some children attended school intermittently

or were away for short periods with their families.

Perso nal D et ai ls (name s/age s)

Age was a basic demographic detailwhich was usually easily obtained. However a proportion

of older Aboriginal people did not know their age. lf the population list was not available the

age woutd be guessed, with a correclion occurring at the data checking stage' A check of

lorty such data sheets revealed wide variation in the accuracy of the predicted age, based on

appearance or "detective work". The biggest difference was twenty five years w¡th a mean

difference of -/+ six years. The accuracy of the population list was assumed for all people born

from 1935 on, lhe period of mission setttement. Ages for people prior to this must have been

guessed by the early mission statf based on the physical appearance of the then child or

young adult. The oldest person examined was 89 years of age.

218 Chapter 5

sheets As described in the methodology section, names were recorded on the examination (1) phonetic vs actual and usually crossed off a population list. Some ditficulties arose with spelting of names, eg Charlie Tjampu being recorded as Charlie Jumbo by survey team members. This was overcome by allowing people to write their own names or obtaining the problems with assistance of a health worker. Correct spellings and identification avoided later the data sheets. (2) Another problem arose with data sheets of people having similar (3) Lastly sounding names, sometimes the sheets were incorrectly assigned as duplications.

nicknames and "while" vs Aboriginal names often caused confusion and only with efrors experience, and the use of the population lists to cross-check people's names, were

avoided

Finally, the Anangu c-ustom ol changing their name when a person sharing either a Chrilian or

surname died required both sensitivity and some detective work, because the name was rþt person allowed to be spoken. For example, if someone in the community called John died, a

named John Smith would become Kumana Smith. After a period of time they would take a

new name and become Jim Smilh, the name John would usually not be spoken, allhough a

name could be pointed out on a populat¡on list. An extension of this custom would be lo both

names, e.g. Kumana Kumana, which provided no direct clue to the survey team of the

persons former name as recorded on the population list.

Language

Atthough most Aborigina! people who participated in the survey spoke English, a lew,

particularly older individuals, relied on relatives or health workers to interpret for them. For the

simple questions asked ¡n the survey (eg. age, sex, name), and the simple tasks pelormed

(e.g. using the Snellen E chart), no problems were encountered. However, more detailed

histories or examination were occasionally ditficult because of the language baíier. Difficulty years in the way ideas are expressed could also lead to misunderstandings. From many of

interpreting in Pitjantjatjara, Edwards (1990), has found that asking questions in the negative

form can lead to conlusion as to the correct yegno answer. This has been called the 'yes

219 Chapter 5 person" syndrome (Mobbs 1986). Confusion also exists because comparisons are not expressed easily in Aboriginal language, e.g. bigger or better may not be easy to ¡nterpret

(Edrruards 1990).

220 Chapter 5

5.2.2. Analysls ol lntra.observer and lnter-observer Error

The lggg/g0 survey utilized the services of two principal investigators, AB being a doctor and

CD being an ophthalmologist. Five sessional ophthalmologists were also employed

(DM,M,ND,RS,DG). The apprOx¡mats proportions of the total population graded for trachoma by each investigalor were

AB - 60% CD -30/"

Dil/l/i/VND/RgDG - 10%

For fhe assessment and diagnosis of visual impairment (when the visual acuity was d36 or worso in the better eye) the s¡x ophthalmologists saw 97% of all individuals, (CD seeing approximately 67 \oof that total) and AB seeing the remaining 3%'

Although standardization of trachoma grading was pelormed using slides, grading manuals

and comparison in the field, no formal inter-observer test¡ng was performed due to the small

numbers being examined by most of the examiners. The two principal researchers (AB and

problems in CD) graded 90% of all examined individuals and conferred regularly on grading

the f¡eld and prior to each trip

Du p licate Ex ami n alí o n Sñeets

To achieve a measure of intra-observer and ¡nter-observer eror, 110 dupl¡cation examination

sheets were randomly collected from those generated by repeat examination on the same

individual (approximately hall the duplicated total). The repeat examinalion had occuned

either on the same visit or on a subsequent visit to lhe same or ditferent community, and was

only discovered during the data checking procedures as outl¡ned in section (3.2.6). As such

they represented a completely unbiased check of interobserver or intra-observer erTor.

Although not evaluated in the literature, formalized testing could have "anticipatory bias'as a

factor. The knowledge of a test biasing the results towards closer agreement because the

221 Chapter 5 participants know they are being evaluated. Because active trachoma could vary over time, scaning could only trachoma scarring and abnormalities of the eye were assessed. Trachoma the eye have arisen over several months but this was felt to be unlikely, and abnormalities of in the would also be unlikely to evolve over a short time span and not be commented on examination sheet.

found for of 110 duplications, 607o were female and 407o were male. concordance was three virtually all visual acuiry testing allowing lor a variation of one line between tests. Only no worse individuals varied more than one line, each lor two ¡¡nes only, all the visions were

than 6/36. Table 5.2.2.1 outlines four eye conditions which were found among the 110 presence individuals, concordance was found between examination for the conditions' or

absence.

Table 5.2.2.1

Number of Examination Condition individuals agreement

Diabetic retinoPathY 3 present or absent cataracts 5 present Comealiniury 3 present Absent globe 3 present

The presence or absence of scarring, trich¡as¡s or comeal opac¡ty was recorded for every

individual. Table 5.2.2.2. ouilines the age distribution, trachoma grading and concordance

and non-concordance between tests. Table 5.2.2.3 compares the grading of trachoma

between the first and second examination.

222 Chapter 5

fable 5.2.2.2.: Concordance and non-concordance in 110 duplicated examinations

Age-Group Concordance Non-concordance

N TS TT/CO TS TT/CO Total

4 0-1 4 2-4 4 4 21 21 5-9 12 1 0-14 11 ; 1 5 1 5-19 4 20-29 10 3 13 30-39 11 2 1 14 40-49 5 4 2 5 16 6 50-59 1 2 2 1 60+ o 3 2 1 15

Total 71 20 7 1f I r10

TT Trachorna trichiasis N = NlorTnal =

CO Cornealopacity TS = Trachoma scaning -

Table 5.2.2.3.: Grading of trachoma: First vs second examination

Second examination

N TS TT/CO

N 7',| 7

First Examination TS 4 20 I

TT/CO 7

223 Chapter 5

were concordant for The analysed data indicated that 98 of 110 or89% of all duplications in that twelve, seven had trachoma status. only 12 of 110 o¡ 11"/owere non-concordant, but One had changed from changed from normalto TS and five had changed from TS to normal' was therefore 2 TS to TS and trichiasis. The effective, or net change in the trachoma scarring in a group with ol 110 or 1.go/o(assuming TT/CO would still be graded as TS). This occurred (aSSuming the non- an overall trachoma prevalence (TS + TT + CO) of 39/110 or 35.5% was calqrlated using the concordance group to all have trachoma signs). A Kappa (K) statistic

following data (Table 5.2-2.41'

Table 5.2.2.4 : Presence or absence of trachoma scaning: Tme 1 vs Ïme 2

ïme 2

Absent Present

Absent 71 7 78 ïme 1

Present 5 27 32

Total 76 34 110

ps = (0.691 x 0.709 + 0.309 x 0.291)

= 0'580

K = (0.891 - 0.580y( 1.0 - 0.580)

=0'74

The value of K (0.74), indicates that the agreement between the dupl¡cate examinations was

good.

224 Chapter 5

5.2.3. Effect of Magnlflcatlon vs Non'Magnlllcatlon

survey To assess the impact of non-magnification on trachoma grading, a separate tesvre-test was performed in Oak Valley during 1990.

examiner and one A total of 116 individuals were examined over a three day period. Only one The recorder were involved. The first examination was perlormed with the naked eye only' second examination, using x 2 loupes, occurred either later on the same day or on a Only subsequent day, the examiner being unaware of the previous examination result. trachoma grading was Performed.

The age and sex distribution is ouilined in Table 5.2.3.1. to indicate the similarity of the Oak

Valley survey to the 1989/90 survey..

Table 5.2.3.1.: Age and sex distribution

tesVre-test at Oak ValleY 1A1U90

Age-Group Females Males Total % of Total

0-1 I 0 1 0.9 2-4 7 2 9 7.8 5-9 2 18 30 25.9 14.7 1 0-14 t 11 17 10 8.6 1 5-19 7 3 12.9 20-29 1 0 5 15 30-39 4 3 7 6.0 40-49 3 I 11 9.5 4.3 50-59 1 4 5 60+ 6 5 11 9.5

Total 57 s9 116 100

The sex distribution was more equal than in the 1989190 survey sample' However, the age population distribution was quite similar (section 3.2.1). The prevalence of trachoma in the

given in Table 5.2.3 .2.werc the results of grading by naked eye examination only. Trichiasis

and corneal opacity were not included.

225 Chapter 5

Table 5.2.3.2.:Trachoma status of individuals in Oak Valley September 1990

naked eye evaluation onlY

Nonnal Trachoma Trachoma Trachoma follicles scarring intense

F M F M F M F M Age-group

0-1 1 0 2-4 3 2 ; 5-9 I 1 0 2 ; ; 0-14 4 7 2 4 15-19 7 3 20-29 6 3 ; ; ; 30-39 4 1 2 40-49 2 4 1 1 3 1 50-59 1 2 1 60+ 0 2 5 3 1

Totals 3 6 34 I 13 11 11 2 I

Active inllammatory trachoma was found in24 (20.7%) individuals, predominantly in the urder

15 age-group. Trachoma scarring was lound in22(18.9o/o) of individuals predominantly in the

20+ age-groups. Table 5.2.3.3 outlines the changes in trachoma grading in going from naked eye to loupe examination.

226 Chapter 5

Table 5.2.3.3.: Trachoma grading, naked eye vs

x2 louPe examination.

Naked eye examination

1 N TF T TS TT CO Com

Loupe examination

N 70 2

TF 4 14 23 T TS 3 13 TT 1 co 1 2 Com 1 4

¡ntense scarring. 1 . Com is any combination of trachoma lollicles, trachoma orlrachoma

2 Trachoma scarring and trachoma follicles.

3 (1) One case of lrachoma scardng and trachoma follicles.

(2) One case of trachoma intense and trachoma follicles'

Concordance in the trachoma grading occurred in 104 individuals in going from naked eye to

x2 loupe examination. A Kappa stat¡stic was calculated using the following data (Table

5.2.3.4).

227 Chapter 5

Table 5.4.3.4: Trachoma grading naked eye vs x2 magnification

x2 Magnilication

Trachoma Trachoma Total Absent Present

Trachoma 70 4 74 Absent Naked eye Trachoma I 4 42 Present

Total 78 38 106

Pe (0.736 x 0.698 + 0.358 x 0.396)

0.655

K (0.e8 - 0.6s5)/(1 -0.6ss)

0.942

The vatue of K (0.942) indicates that there was excellent agreement between naked eye and

x2 loupe examination.

It appeared that although the trachoma grading changed in 12 (10.6%) individuals, the

changes resulted in a net difference of one in trachoma scarring and ol two in inflammatory

trachoma (either TF or Tl). Relating these ditferences to the trachoma data'or naked eye

grading (Table 5.2.9.21, a net change of 1 in a totalof 22 individuals represented an enor of

4.So/" and a net change ol2 in a total ol 24 individuats with TF or Tl represented an enor of

g.4%. Both changes were towards normality, indicating that naked eye examination could

produce a higher prevalence of inflammatory or Scarring in trachoma grading compared to

loupe examination.

228 Chapter 5

Summary ln the lggg/g0 survey an analysis of inter-observer and intra-observer error was undertaken using duplicate examination sheets. The data was grouped together so inter-observer and good intra-observer comparisons could not be made separately, The results demonstrated

eye agreement in trachoma grading ol TS, TT,and CO. The results for visual acuity and maior pathology were almost unchanged between examinalions'

The results are applicable to the general survey ¡f it ¡s assumed that those who unden¡rent a presented for the second examination are representative of the general population' The data age distribution, prevalence of trachoma and eye disease is consistenl with this hypothesis'

Therefore, there is no reason to believe that those individuals seen twice have different characteristics from those presenting only once.

A study comparing magnilication (x2 loupes) with non-magnificat¡on demonstrated excellent

agreement between examinations. tt was also concluded that non-magnification tended if

anything, to over estimate the prevalence of inflammatory trachoma.

229 Chapter 5

5.3 Methodotoglcal Problems ln the Survey and wlth the Comparlsons

presented in section 1'4' The methodological problems associated with eye surveys was

Some issues ol specific concern to the 1989/90 survey were:

The Number of Subiects

from the The total number of people surveyed was more than that required, as calculated survey was sample size (section 3.2.6.). There were several reasons why a complete (R'A'C'O' attempted. A comparison of the resutts with the N.T.E.H.P. survey 1976-1979

Report 1gg0) was to be made, and it was felt that the study design should be similar.

Secondly, a simple randomized survey could not be effectively conducted because

population lists were not known to exist when the design ol the survey was completed'

Thirdly, the survey was to have a service role and individuals could not be refused assessment for and treatment. The larger sample was fortuitous because the odginal number calculated

trachoma did not include an ad¡ustment for household clustering. A fac{or that has been

recenily identified as necessitating an increase in sample size to ma¡ntain a reasonable

variance of prevalerrce estimates (Katz et al. 1988).

generated from Several problems could arise ¡n the conduct of a survey. A reliance on dala

short visits to each community, or from only one or two communities could result in selection

bias. On[ fhose w¡th eye problems may have presented to be examined, or if several were hundred people had been seen in one community during a lootballweekend those who

blind may have been excluded. All communities needed to be included, as there was A variability in both participation rates and disease prevalence between them (section 3'3).

study comparing self-selection and random sampling methods has dernonstrated an under' the estimation of btindness of 130%, suggesting self-presentation may grossly underestimats

prevalence of blindness (Said et al. 1973). To avoid this problem a pro'ac{ive role was

adopted by the survey team to encourage participation by as many people as possible' An participat¡on rates attempt was made to sample evenly across al! age-groups, although the

(section g.g.2.) demonstrate the difficulty in achieving this. Although only a 54olo participation

230 Chapter 5

proiected 1986 rate was achieved overall, based on the population list denominator, using in each census dala the figure was 76%. lt was the impression of the research team that when community, at the time ol the v¡s¡t, up to 90% of the population was examined'

Ernabella was visited in 1990, 1Oolo ol the population was ¡dentif¡ed as being away permanen¡y or semi-permanently (section 3.3.2). Many others, usually ¡n the 20-39 age' group were also absent. Therefore, the study was probably as successful as it could be in obtaining a comPlete survey.

Within the limitations described above, those who were blind may have chosen not to be examined or because of diffidence or shame not presented to the survey team. The high response from the 60+ age-group, which had a high prevalence visual impairment, could lead to a hlgher estimate for btindness if the sample population only was used lo generate prevalence rates.

Disease Defi nitions and Classifications

A review of the methodological problems of the otd trachoma grading system (section 4.2.)

and the apparen¡y good intra-observer and inter-observer agreement with the new 5 sign

system, has already been presented (sec{ion2.1.41. The standardization procedures forthe

eye survey in this thesis have been outlined (section 3.2.6). These were similar to other

surveys which had equivatent methodology. At the time it was felt unnecessary to test for

intra-observer and inlerobserver agreement because the l¡terature had already demonstrated

good agreement between graders with the new 5 s¡gn system (Taylor et al. 1987 c ).

Retrospectively, duplicate examination sheets were examined lo test for intra-observer and

inter-observer agreement for grading trachoma scarring and visual acuity testing (seaion

5.2.2.1. Formal testing of the grading for trachoma follicles and trachoma ¡rìtense perhaps

shoutd have been undertaken, part¡cularly because of the difficulty with the grading of Tl in

the older age-group (section 9.4.21. ln previous surveys ¡t seems that although adjustment of

the final prevalence rales based on the test results was not performed, some benefit was

presumably derived from the process to help standardize the results between graders. A

231 Chapter 5 process in the 1989/90 survey which was achieved by the principal graders, AB and CD, by their review of slides and consultation in the f¡eld.

Ophthalmic Measures

Visual acuity test¡ng was performed under varying conditions inside and outside clinics, with and without other people present. However, the subject was always at six metres and care was taken to avoid prompt¡ng. Classification of individuals ir¡to good vision and blindness was retarively quick and was also accurate. People who had corrected visual acuity of 6/18 or worse in the better eye were re-checked once by the f¡rst examiner and secondly by an

ophthalmologist. Retrospective examination of duplicate examination sheets showed good

reproducibility between testing. Another method of checking results could have been

employed. As used in a larger survey in Egypt, a random sample of screened individuals wlto

had "good vision" could have been re-examined to check for false negatives, those who had

visual impairment but who were not identified as such (Said et al. 1970). A process achieved

in the 1989/90 survey with the duplicate examination sheels.

I /ithin the constraints set by the local environment and the ability to communicate with the

participants, the results for visual acuity obtained would have accurately classifìed individuals

¡nto the four categories of visual acuity mentioned in the results (section 3.3.6), some vadation

may have occurred within these groups, but as demonstrated by the review of duplicate

examination sheets this was only a minor problem. Other studies have dernonstrated good

inter-observer agreement in field survey work for visual acu¡ly, with Kappa's between 0.80 and

0.8g (Briiliant et al. 1983 b ). However, in clinic settings statistically significant variations have

occurred in visual acuity data coltec{ed by clinicians and nurses for patients with cataracts

(Gibson and Sanderson 1980).

One area for which standardization was difficult to achieve was the cause of visual impairment.

The small size of the survey limited the extent to which lormaltesting of the ophthalmologists

could be achieved for the range of conditions seen. This problem has been observed in

studies designed to assess inter-observer agreement, with adequate evaluation impossible

232 Chapter 5 for rare causes of úsualimpairment and variable agreement for other more common causes of blindness (Brill¡ant et al. 1983 b ). Even surveys that have emphasised quality assurance e.g'

The Framingham Eye Survey, had problems with diagnostic standardization, with considerable variation between examiners (Kahn et al. 1975). Several problems could añse.

Firsily, when determining the principal cause of blindness there may have been several conditions affeAing the eye at the same time. ln the 1989/90 survey for example, the most usual being trachoma and cataract in combination. secondly, trachomatous corneal opacity could co-exist with corneal opacity from olher causes e.g. labrador keratopathy. Although it is

possible to distinguish corneal opacity due to trachoma from that due labrador keratopathy

(McGuiness et al. 1972), when the two conditions coexisted, the attribution of blindness to a

particular cause was diflicult. Additionally when non-trachomatous corneal opacity was

identified the underlying cause was sometimes ditficult to identify; medical rec,ords did not

necessarily document the original injury or infect¡on, and many initiating events occurred in

childhood; the passage of t¡me blurring the memory of the true cause. Thirdly, it was also

possible for one condition to obscure the diagnosis of another; for example, a treatable

cataract might obscure an untreatable optic atrophy or retinal detachment. Therefore,

although many calaract extractions could be expected to relieve blindness, in some ¡nstances

a secondary condition might prectude the restoration ol good vision, leading to

disappointment for both the individual and the treating ophthalmologist'

Comparison of Results

The methodological problems associated with comparing the 1989/90 survey with the 197&

79 N.T.E.H.p. survey (R.A.C.O. Report 1980) has already been discussed (sect¡on 4.2).

Two factors which would be expected to increase the accuracy of the comparison were; the

same communities were visited in both surveys, with the populat¡ons examined having similar

age structures, secondly the study designs were similar. However, the 1976 survey probably

had better pre-visit organisation than the 1989/90 survey, although the latter built upon the

previous lrachoma program visits. Other ditferences included the amount of time spent in

233 Chapter 5

more each community, and the time span of the total survey. The 1989/90 survey spending time in each community and us¡ng two years to complete the survey work.

The extens¡on of the survey over two years could theoretically lead to some methodological problems. 1. Differences in the weather might affect the prevalence of trachoma, e'9. via the prevalence of flies. 2. The occurrence of ceremonies, with the influx of a large group of

people, could have produced an epidemic ol trachoma or other infective coniunctivitis eg. an and was outbreak of gonococcal coniunctivitis was followed across Central Australia in 1991

linked to the movement of people associated with ceremonial ac{ivity (Condon 1991)' 3.

people could more easily be double counted in the same or different settlement. The data

checking procedure in fact identified over 200 duplications atter examination and several

dozen before the second examination took place. 4. The etfect of seeing and treating many

individuals in the first year may have altered the prevalence of disease in the second, with the

greater awareness ol trachoma and eye health that the first round of examinations

engendered. The awareness leading to greater setf-identification of disease or greater case

finding by health workers. A mitigating feature of the 1989190 survey in this regard was its

resembtance to the usual South Austratian Abodginal Trachoma Program trips which had

visited the communilies once or twice a year during the 1980s. The survey therefore

rellected the usual pattern of visits.

It may be concluded that atthough a prevalence survey of trachoma is a'snap shot'or a cross-

section in time of the ac{ual eye health picture, the degree to which the length of a survey; be

it 2 weeks, 2 months or 2 years, creates a different result is ditlicult to assess' Certainly il

significant interventions occur between the start and linish of a survey the results may be

biased; however, if the environmental features, socioeconomic factors and medical facilities of

the communities remained the same, there would be no reason to suspect that the results of a

2 week survey would be and ditferent from that taking 2 years as long as thè coverage of the

population was the same.

234 Chapter 5

the one of the hardest methodological problems to resolve was the choice between using The sample sample or source population as a denominator lor prevalence rate calculations. would be subiect to bias, because relatively, the older age-groups were oversampled' years increasing the crude prevalence of visual impairment, and the younger age-groups 0-15 were also disproportionately represenled, possibly leading to an over estimation of crude greatly atfected, trachoma prevalence. Although age-specific prevalence rates would not be presentation of crude proportions could give inaccr¡rate results for the population as a whole' given. The choice A simple random survey could not be conducted for the reasons already then became which est¡mate of the source population was most accurate. Both the population lists and projected 1986 census population distribution had flaws. The latter having problems with its collection and not allowing for net migration and the former being subject to double counting between communities and increasing the estimate by not being

regularly updated for people leaving the communities. Whatever the drawbacks, the crude

prevalence rates derived for blindness and poor vision using the population list estimates were thought to be a more accurale reflection of disease than those based solely on the

sample proportions. Again, age-specific rates using the sample denominator would have

reasonable validity, the high sampling in some age-groups (60+ and 0-15) closely

approximating the source population reducing the ditferenco between the estimated rates.

Finally, the advantages and disadvantages of using cross-sectional surveys to monitor change

in a defined community have already been discussed (section 1.4). The surveys used in this

thesis have been population based, have had similar survey design and study procedures,

have studied diseases or ophthalmic measures with standafd¡zed definitions, and the

diseases studied do not have large swings in prevatence. Therefore, like comparisons in

cardiovascular disease (Dobson 1987), the results of the compadsons are probably valid.

235 Chapter 5

5.4 Ethlcal lssues: Research ln Aborlglnal Communltles

Several ethical considerations were identilied prior to lhe survey, these were addressed in submissions to the four ethical bodies consulted. ln addition, ethical problems arose during the survey and after the conclusion of the field work. Many of the issues wera not specific to

Aboriginal health and research, and relate to the conduct of any research in a definable

population. These issues included:

i. The risks for researchers and trachoma program personnel in travelling to rerþte

areas. Additionally, there were risks for the subiects to be screened in travelling to tho

examination centres. The eye examination itself had minimal risks associated with it. Every

effort was made to ensure the cleanliness and safety of the equipmer¡t used.

2. The potential disruption to a small community caused by a survey or research proiec{

with its influx of people and equipment. However, allthe communities visited had previous

Trachoma Program visits, and the survey work was inlegrated into the normal service provision

of the Trachoma Program, as such the survey was well received by the comrnunit¡es.

The eye examination was non-invasive and part¡cipation d¡d not appear to be atfected by the

gender of the examiners (it was gender neutral). Furlher, personal data were kept to a

minimum to avoid alienating participants.

g. lnformed verbal consent was obtained from all adults or parerûs of parlicipating

children. Written consents proved to be unwieldy for lield use, and unnecessary for the

simple examination performed. The administration of eye drcps for diagrrcsis or treatment was

govemed by the normal medicalconsent procedures of a Trachoma Program trip.

4. There was no coercion in obtaining community or individual consent, their lreedom ol

choice was respected at alltimes. Although the part¡cipating communities from the A P lands

were represented on the Nganampa Health Council, (which reviewed the research proposal),

additional meet¡ngs often took place prior to the commencement of work, between the

communily council and the medical research personnel. The extent and benefits of the

236 Chapter 5 survey were ou¡ined, and its methodology explained. lnformalion was received from the community concerning the social and cultural imperatives which were important, and questions were answered.

5. The confidentiality of both the communities and individuals was maintained during and after the survey. Although names appeared on data sheets, personal information was known only to the researchers. lndividual clinical details (e.9. for treatment puçoses) were only given to the heafth workers for each respective community. Aggregated data only weie given to community councils, personalinformation was not released.

6. At alltimes the wellare of the indiv¡dualparticipants and communities were paramount

in any conflicl arising from the research. For example, tñps were scheduled to avoid conflicts

with business or ceremonies when this was known in advance.

During the survey several issues arose which had ethical ramifications.

1. The use of medical records had not been specifically requested in the original

submission lor ethical approval, although it was implied from the integration of the service and

research roles. The principal investigator therefore had two responsibilities, the first to the

research project, which had not specifically requested the use of medical records, and

secondly to the Trachoma Program as a doctor. ln that role he would normally, and ethically,

utilize the medical records to obtain and update information'

It was only on reflection by the principal investigator that the use of records for people not

seen (to identify blind individuals) was seen to be unethical in terms of the original submission'

However, it had been justilied as part of the Trachoma Program work, which sought to identify

those in need of trealment and follow up.

2. During the survey, the research team became aware of several socio-cultural

constraints to the research process. When a death occr¡ned within a community prior to a visit,

many people would leave the immediate area. Although some individuals were happy to

participate in the survey the wishes of others not to participate were respected. Similarly, with

237 Chapter 5 ceremonies or "business', only il the community was happy for the survey to continue, did the

heatth survey work progress. On a personal level, the research team respected the wishes of workers either not to be involved at all, or not to approach cerlain members of the community, to whom, according to Aboriginal law they could not speak (avoidance relationships).

Additionally, the family/community structure and the role of parents or guardians was respected at alltimes. This extended to the role of the trad¡tional Aboriginal medicine men or

Ngangkere, whose healing powers in Aboriginal society were acknowledged by the survey team.

3. The service role of the research has been previously described. This was extended

during the survey period to include workshops in eye care and trealment for Aboriginal health

workers. Skill sharing occurred with the health wod

Discussions were also held with community members on strategies to improve eye health

using a primary health care approach.

4. No problems arose from the type of physicat examination (eversion of the eyelids)

pelormed. Most individuals wers happy to be examined in front of others; however, some

prefened to be tested privately. This was respected and provision for it was generally available

in most locations.

5. The survey team comprised of two or three, and sometimes four individuals. Because

of the teams small size, the exploitation ol community resources was .not a major issue.

Usually an area of the medical clinic was used which avoided disrupting the day-to-day routine

of the service. The survey leam utilized visitors' accommodation for which a fee was payable

(often waived), but sometimes research personnel shared or were given quarters used by the

medical service statf. This occurred not from necessity but friendship. Food was purchased

from the local community store, thus in a small way, the researchers supported the local

economy.

238 Chapter 5

Atthough community health workers were diverted from their usual roles to assist the survey team, the strong service focus ensured that the health return for the t¡me spent was worthwhile for both the health service and the community.

At the conclusion of the survey two issues were important. i. The ownership ol the data and the publication of research material was an ethical issue of great importance to the Aboriginal people.

During the survey intefim reports were given to the communit¡es. At its conclusion a summary report was given to Nganampa Health Council (N.H.C.) ol data for the Anangu Pitiantiatiara lands. Data from Yalata were not included. After considerat¡on by N.H.C. the report was fon¡varded to the A.H.O., the Trachoma Program, and the director of the Yalata Health

Services.

A final report which detailed all the resuJts and the comparisons with the 1976 survey was prepared and given ¡n the same sequence to these organisations.

The N.H.C. saw the data and conclusions first, to allow approvalto be given for its furlher

distribution. A co-ownership model was developed to ensure that the form and presentation

of the data was acceptabte to both the comrnunities from whom the data were derived, and the

researchers who collected it. lt is hoped that data for publication will also be handled in the

same way, ensuring that the interests of the Aboriginal people are respected and that the

distribution of public information is by mutualagreement'

2. The use of research material beyond the research process will occur for the original

data sheets. Although not specifically outlined in the original submission to the ethical bodies

it was mutually agreed that the data sheets eventually be retumed to the communities after

use by the Trachoma Program. The Trachoma Program would utilize transcribed sheets as

individual medical records,lor recalland lollow-up purposes. The data sheets would then be

returned to the communities for incorporation into the clinic medical records.

239 Chapter 5

5.5 Dlscusslon

This chapter has examined three principal research issues which related primarily to the conduct ol the eye health survey and the evaluation of comparisons with previous work.

However, ol particular importance were the ethical issues associated with research in

Aboriginal communities and the specific issues faced by the the 1989/90 survey in working with the Anangu Pitiantjat¡ara.

The practical problems encountered in the 1989/90 survey have not been documented in previous eye studies, and so comparisons can not be made. However, cr¡lturalfaclors and the implementation ol western medicine have been discussed by some authors. Firstly,

Maddocks and Maddocks (1991), found that in a Papuan village population, westem ideas and structures, including medical concepts, could depdve people of the control they once assumed. The relevance of any intervention should therefore be measured by the degree of control the particular society maintains. Knowledge and autonomy should be encouraged.

For the Anangu Pitjantjatjara some control is already exercised by the Abodginal controlled

health service. But, we still do not understand the Anangu concept of eye disease

(particularly trachoma). lt seems that trachoma is viewed as a'white man's" disease, and

therefore treatment with antibiotics is accepted. However, socio-q¡ltural attitudes to hyg¡ene

(e.g. face washing) is less clear and anecdotally there was some resistance to lhese measures

because they were 'rarhite man's" ways. This superficial analysis requires further exploration

and could lead to a greater understanding of Aboriginal culture.

Secondly, attitudes to education and hospitalisation have been identified as baniers to

accessing health care (Watson 1987). However, the involvement of Pitiar¡tjatiara women in

the management of chronic diseases such as diabetes and hyperlípidemia has been achieved

when they were given autonomy (Mulvey and Endean 1991).

Thirdly, the whole concept of western medicine and its application to Aboriginal people has

been questioned. Medical treatment is thought to reflect and reproduce the cultural,

socioeconomic and politicalforces of a wider society. Additionally, westem medicalpractice is

240 Chapter 5 seen as particularistic, biophysical, mechanistic and disease orientated, whereas Aboriginal health beliefs and practices are holistic (Nathan and Japanangka 1983). However, although the gulf between them is cutturally wide simple changes have ocq¡rred and the acceptance of western treatments indicates an evolution in both European and Aboriginalthinking.

The analysis of observer error demonstrated that the simptification ol the f¡ve sign system lor trachoma grading ensured good agreement between observers or observations. The use of duplicate examination sheets has not been described before, although descriptions ol standardization procedures occur in the literature as do evaluation of intra-observer and ¡nter- observer errors. The feature of both these procedures is the possibility of "anticipatory bias', where those being tested know that an evaluation is taking place. They would then increase their effort to maximize their score, this maximal etfort could not be susta¡ned throughout a

survey, and therefore the evaluation can not be truly representative of the survey' The

method used during the 1989/90 survey was an evaluation of duplicate examinations that

occurred by chance. The graders and subjects were unaware that a particular examination

would be subsequently used for intra-observer and inter-observer error, freeing the test from

any .ant¡cipalory bias".

The effect of non-magnification on trachoma grading has again not been descdbed in the

literature before, and the evaluation was undertaken because it was observed in the field that

some health workers didn't use loupes when assessing trachoma. Rather than erforce the

use of loupes lor the clinical assessment ol trachoma, the evaluation of the¡r uso

demonstrated that under test cond¡tions no significant differences were observed between

magnification and non-magnification. Health workers could therefore be confident that the¡r

diagnosis would not be atfec{ed by not using loupes.

The methodological problems encounlered in the survey were similar to those encountered

¡n other surveys; however, population lists have not been used in Australia before to def¡ne a

population (although the¡r use overseas is documented (Thelefors 1987)). Their use

contributed to the accuracy of the 1989/90 survey. The simplification of disease definitions

made the trachoma grading easier and although comparisons with previous surveys was more

diflicult, satisfactory adjustments could be made for inflammatory trachoma. Comparison for

241 Chapter 5 trachoma scarring was not performed because it was fett that the 1989190 survey scored TS grade I (N.T.E.H.P. review definition) much more than the 1976-1979 survey , and the results obtainabte for that survey (from the 1985 review) were lor TS grade 2+(ie did not ¡nclude TS grade 1).

Finally, there were several ethical issues raised during the course of the survey' Medical and public health research invotving Aborigínalcommunities has undergone important changes in the last five years (Houston and Legge 1992). The adoption of the National Aboñgina! Health

Strategy (National Aboriginal Heatth Strategy 1989), and the publication by the National Health and Medical Research Council (N.H.M.R.C.) of a slatement on ethics in Aboriginal research

(N.H.M.R.C. 1992), have highlighted the need to work with Aboriginalcommunities in solving

issues identi¡ed by Aboriginal peopte, and not, as has occurred in the past, performing

research on Aboriginal people. These guidelines also encourage the use of informed

consent lrom the group as well as the individual and suggest that a high level of control by the

group in the conducl of the research, or use of the results, is a worthwhile feature (Maddocks

199i). Community participation should encompass setting priorities for research, ioint

ownership of the project, involvement in the actual process, and the right to withdraw if there

is a break of social, q¡ltural or proiect protocol (Wyatt 1991).

The 1989/90 eye health survey comptied with most of the guideline's set down in the report

ot the Nationat Health Strategy Working Party (1989), recommendations 11.1.1. to 11.1.9..

Only in one area did the study not involve the community, the initialformulation of the study'

However, the Trachoma Prcgram was involved very early in the planning Process.

242 Chapler 6

Chapter 6. Conclusions and lmpllcatlons

243 Chapter 6

6.1 lntroductlon

The unacceptable health status of Aboriginal people in South Australia has been documented previously (Easterman et al. 1988). On a range of general indices, including, education, employment status, and economic status, the Aboriginal population is disadvantaged compared to non-Aboriginals. Further, health indicalors such as: fert¡lity, mortality, and rates of hospitalization, are all substantially worse in Aboriginal people than in non-Aboriginal people (Thompson and Briscoe 1991).

Contemporary issues in the health of Aboriginal people have been documented recently

(Tozillo and Kerr 1991), and reflect not only specific diseases, but general issues such as: management and planning in community controlled services, environmental health, the provision of health hardrryare and the use of health promotion. Although lrachoma is one of many diseases, the approach to ils control can be used as a model for other problems, incorporaling, as it does, primary through tertiary levels of intervention.

244 Chapter 6

6.2 Dlscusslon

The results of the 1989/90 survey have several implications. From its genesis as a population based prevalence survey of trachoma and visual disability in an Aboriginal population, the research developed into a review ol the historical and epidemiological aspects ol the eye health in the Anangu Pitiantjatjara.

The literature review revealed that the Anangu Pitiantiatjara have been the sub¡ect of several pubtished trachoma and eye health surveys since the 1950s. Additional information relating to the eye health of the Aboriginal population had been noted in anthropological and exploratory expeditions since the 1880s.

Assessment of the avaitable literature and the results of all the published survey work, including the 198gl9o survey, suggest a changing pattem of disease consistent with the late introduction ol trachoma into the north-west of South Australia. lts initial absence in Gentral

Austratia does not imply that Aborigina! people in the rest of Australia were not afflicted by trachoma before the coming of white man. Trachoma may have been endemic in some areas but its carriage and transmission in Central Australia, amongst a people who knew few other infectious diseases and lived in small nomadic groups, would not have favoured an organism that thrives in crowded, unhygienic conditions.

The review of the literature relating to trachoma demonstrated the comprehensive but

incomplete knowledge about both preventive strategies and treatment programs (section 1.3

(2.1.8.-2.1.9). For example, hygiene praclices such as face washing were identilied as an

¡ntervent¡on strategy to reduce lrachoma (Taylor et al. 1989); however, further stud¡es

suggest that cleaning nasal discharge and keeping flies otf children's faces was rþre

important (West et al. 1991). Apptication of face washing to an Australian Aboriginal

community had no effect on the prevalence ol trachoma (Peach et al. 1987). The differing

results indicate that although overseas research could suggest areas for primary prevention,

local experience and evaluation are necessary to just¡fy q¡llural and other changes.

245 Chapter 6

The treatment of trachoma in the individual using either oral or topical therapy is well

documented and accepted. The effectiveness of treatment on a mass scale is less clear

(section 1.g (2.2.9) ). The use of a campaign style approach is likely to meet some polit¡cal

resistance if attempted on a scale which would have a benelicial etlect. Nevertheless, lhe

application of a culturally appropriate and systematic treatment program, in a group of

communities served by one health service (e.9. Nganampa Health Service) could be

successful, as demonstrated by overseas experience. Such a strategy has already been

proposed in the Northern Territory by the Menzies School of Health Research (Matthsws

1987). Reintroduction of trachoma from other communities would occur but the chronic cycle

of closely spaced repeated ¡nfect¡on could be broken, leading to a reduction ln the severity if

not the prevalence of trachomatous disease. Another approach could utilize systematic case

finding and treatment of families by Aboriginal health workers. This should have been

occurring; however, ¡t appears that the demands of multiple problems had limited the ability of

the health services to resPond.

The last implication, and perhaps the most important of the l¡terature review was that

socioeconomic development, infrastructure improvement and some attendant cultural

change was the most effect¡ve means of decreasing or eliminating trachoma (section 2.2.n'

However, to wait for this to occur in rural Aboriginal communities could condemn many

individuals to years of blindness.

The results of the eye health survey presented in chapter 3 indicate thal trachoma is still

endemic among the Anangu Pitiantiatjara of South Australia, and that whilst visual acuity is

good in the younger age-groups, a considerable burden ol blindness already exists in the old.

It appears, although this is by no means certain, that these younger age-groups, as they get

older, will not progress lo the same levels of visual disability as their rnothers and fathers.

Trachoma is an indication of socio-economic development, community and health

infrastructure (Marx 1989); that trachoma still exists demonstrates that despite positive

changes within the communit¡es, improvement can still be made. For examplo, an

246 Chapter 6 improvement in the prevalence of follicular trachoma in Western Auslralia was found to be associated with improvements in school hygiene and housing (Cooper et al. 1986).

Socioeconomic development would not only benefit eye health but would also address the identified matrix of disease which had been documefied previously (Hollows 1982)

The high prevalence ol blindness has implications for the communities beyond the immediate provision of treatment services. There is a personal cost to the individual and costs to family for the additional care of a blind person. Much of the blindness ident¡f¡ed in the 1989/90 survey was either preventabte as with trachoma or easily remediable as with cataracts. A backlog of blindness from catarac,ts does occur in developed countries (Drummond and Yates

1gg1); however, the high prevalence of sight limiting cataracts found in tho 1989/90 suruey, indicates some deficiencies in the provision of treatment services. Either the Trachoma program has been unable to encourage attendance at its clinics and ensure completed treatment, or the local health services have had an uncoordinated approach to the identification and follow-up of individuals with eye health problems. Again, the ditficulties of servicing a remote region which has signilicanl socioeconomic disadvantage, and a broad spectrum of ill health, determines the distribution of finite resources.

The results also demonstrate that more eye disease is found in females; as measured by sample prevalence rates of cataracts and trachoma trichiasis. Other surveys have found similar differences between the sexes (section 3.4.3) but an adequate explanation has not been found, differential access to lreatment may be the most important factor. A biological

explanation for the difference based on sex has not been expounded, and sex therefore may

be a confounding variable.

For the Anangu Pitjantjatjara, perceived barriers ¡n treatment services for women need to be

reduced and socio-cuttural expectations examined. Possible problems could arise from the

perception of invasive procedures being carried out by male doctors; it is reported that the

attendance by women on men can be difficult, and it is preferable for women to treat women

and men to treat men (Edwards 1990). There may also be cultural expectations conceming

247 Chapter 6 the rote of male and female elders vis. the subsequent need for good eyesight. lt has been lound in other surveys that in traditional societies, sex roles have implicat¡ons for knowledge

about access to health resources (Brilliant and Brilliant 1985), and this may require lurther

exploration in Anangu Pitjantjat¡ara culture'

The comparison of the 1989/90 survey with the results of the R.A.C.O. Report of 1980

demonstrate a statistically significant decline in both inflammalory and cicatricialtrachoma. Two

recent evaluations of trachoma in CentralAustralia suggested that in some coastalAboriginal

communities the prevalence of lrachoma has been reduced, whilst it has remained the same

in inland communities (Freeman et al 1985; Meredith et al 1989). The 1989/90 survey

therefore demonstrates for the first time that the eye health ol Aboriginal people in Central

Australia may be improving. However, isolated reports of gonococcalconiunctivit¡s ¡n Central

Australia (Brennen et al. 1989), and more recently in the Anangu Pitjantjatjara lands (Dr'

Mutvany personal communication 1991) indicate that eye hygiene still needs to be improved'

The findings of the survey and analysis require explanation. What changes have occurred in

the communities in the last 15 years? Have the changes reduced the prevalence of other

diseases? What further positive changes can be initiated to c¡ntinue the ¡mprovemenl of eye

health? The changes that might have occurred may be an improvement in water supplies,

nutrition or sheller, the ¡nfluence of an Abodginal controlled medical service, the intervention

of the Trachoma Program or simply hygiene changes. A retrospeclive study would be

methodologically diflicult, but prospective cohort studies, although expensive, could yield

valuable and accurate data if conducted properly (Rosenthal 1980).

The 1989/90 survey identified two areas of the study design which may have ¡mpl¡cations for

future studies of trachoma and eye health. The first related to the standardization and enor

assessment of grading trachoma or testing vision. ln future surveys repeat examination could

be incorporated randomly into the study design to evaluate both intra-observer and inter'

observer error. The second related to the grading of trachoma using x 2 loupes. A small study

of 1tO individuals demonstrated no sígnificant ditferences between grading and TF and TS

248 Chapter 6 both with and without magnification. Atthough not in accordance w¡th W.H.O. guide-lines it demonstrated that grading could occur without magnification with results being comparable to that obtained with magnification'

The success of the survey in seeing a large proportion of the population and a willingness of the communities to participate indicate that a prominent service role should if possible be

incorporated in future studies. The review of ethical considerations reinforces the co-

ownership model for research within Aboriginal communities. Participation by Aboriginal

people in the development, implementation and distribution of the results is important.

Aboriginal people should have the right to review work and be able to suggest appropriate

changes.

6.3 Future Dlrectlons

Although Australia cannot be classified as a developing country the resources available to

Aboriginal communities are limited. lt has been suggested that for areas of limited means

evaluation of personat and family faclors should be undertaken to identify risk factors which

are amenable to intervention (Telsh et al. 1988)'

Risk factors for inflammatory trachomatous disease have been identified in many studies

(Tielsh et al. 1988). Consistently and of relevance to the conditions encountered in the

Anangu Pitjantjatjara and Yalata lands are socioeconomic status, distance to primary water

source, presence of lrachoma among siblings and presence of a latrine. These conditions

have been identified and given priority on the A P lands (Nganampa Health Council 1989).

ln Africa a shortage of ophthalnþlogists has been identified (Foster 1987), and the solution to

high rates of cataract blindness found in various surveys in Africa was to lrain more

ophthalmologists or medical assistants who were capable of perlorming surgery (Faal et al.

1989). By contrast, even though the per capila number of ophthalmologists in England is not

very dilferent from that in lndia or China, the level of access and affordab¡l¡ty ¡s important in

249 Chapter 6

determining who gets operated on for cataract in those countries (Sommer 1988)' The results from this survey indicate that the distribution and nrotivation of skilled practitioners was iust as important as total numbers in providing eye care to disadvantaged commun¡t¡es. The enthusiasm of the Trachoma Program and Community Health Service personnel are essential in developing better access to treatment programs.

When the resutts for trachoma were aggregated it appears that there is a high rate ol act¡ve inflammatory trachoma in the younger age-groups with a high rate ol cicatricialtrachoma in the older age-groups. Superficially this would fit Jones' epidemiological classilicat¡on of a class 1 region, an area of blinding hyperendemic trachoma (Jones 1975). However, this definition applies to a steady state situation and with the demonstrated decline over 15 years, the situation in this study appears to be in transition, and a class 2 region (non-btinding trachoma) appears to be emerging.

250 References

Llst of Relerences

Abbie A.A. 1960, 'Physicalchanges in Australian Aboriginals consequent upon European

contac{s', Oceania, 31: P140-144'

Abbie A.A. 1976,'The Original Australians', Rigby.

Aboriginal Health Organisation 1988, Renat Suruey Report, Aboriginal Health Organisation,

South Australia.

Abramson J.H. 1991, Cross-sectional studies, ln Oxford Textbook of Public Health,eds

Holland W.W., Detals D., Knox G.,Oxford University Press, Oxford: Vol. 2, p107'120.

Abu El-Asrar 4., Van Den Oord J., Geboes K., Missotten L., Emarah M.H., Desmet V. 1989'

'lmmunopathology of trachomatous conjunctivitis', Br J Ophthalmol ,73:p276-282.

Abu-Jaudeh C. 1953'The Occurrence of lnclusion Bodies in Epithelium of Nasaland Urethral

Mucous membranes of Trachomatous pat¡ents. A Preliminary Reporl', Am J

Ophthalmol, 36: P947'956.

Adala H.S. 1983, 'Ocular lnjuries in Africa', Soc Sci Med,17,(221:p1729-1735.

Al-Rifai K.M.J. 1988, 'Trachoma through H¡story', lnt Ophthalmol,12:p9'14

Aiello L.M., Berrocal J., Davis M.D., Ederer F., Goldberg M.F., Harris J.E., Kilmit C.R.,

Knatterud G.L. 1974,'The Diabetic Retiopathy Study', (letter), Arch Ophthalmol,92:

pl 79-1 80.

Anderson J.D.C., Benttey C.C. 1986, 'Role of community health workers in trachoma control:

Case study from a Somali refugee camp', Trop Doct, þr 16 (2): p6e69.

Anderson J.R. 1939, 'Blindness in pdvate pract¡ce', Med J Aust,2: p680.

Assaad F.4., Maxvvell-Lyons F. 1967, 'systematic observer variation in trachoma studies', Bull

World Health Organ,36: P885-900.

251 References

Assaad F.4., Maxwell-Lyons F., Sundaresan T. 1969, 'Use of local variations in trachoma

endemicity in depicting interplay between socioeconomic conditions and disease',

Bull World Health Organ,41: P181'

Assaad F.A., Sundaresan T., Maxwell-Lyons F. 1971 'The household pattem of trachoma in

Taiwan', BullWorld Health Organ, zl4: p605-615.

Atlas of South Australia 1986, eds. Gritfin T. and McOaskillM., Adelaide South Australia.

Australian Bureau of Statistics 1986, Census Data t986, A.B.S. Adelaide South Australia.

Australian Bureau of Statistics 1988, Aboriginal Bitlhs and Deaths, Review oÍ data quality and

statisticat summary, A.B.S. Adelaide South Australia.

Babalola O.E. 1989, 'Trachoma in Kaduna, Northern Nigeria. Recent observat¡ons on

patterns ol clinical presentation', Trop and Geogr Med,41:. p41-44-

Ballard R.C., Fehler H.G., Fotheringham P.,Sutler E.E., Trehame J.D. 1983, 'Trachoma in

South Africa', Soc Sci Med,17: p1755-1765.

Banks C.N., Hutton W.K. 1981, 'Blindness in NSW. An estimate of the prevalence and some

contribut¡ng causes' Aust J Ophthalmol, Nov, 9(4): p285'288.

Banks J.R., Vanden Driesen G., Stark E. 1985, 'Chlamydia trachomatis in smears from eyes,

ears and throats of children with chronic otitis media', Lancet, AugS: 9278.

Barnes F. 1970, The biology of pre-neolithic man, in SV. Boyden (ed), Iña lmpact of

Civitisation on the Biology of man, Canbena, Australian National University Press.

Basedow H. 1932, 'Diseases of the Australian Aboñgines', J Trop Med Hyg,35:pT7.

Batterbury M., Khaw P.T., Hands R., Elkington A.R. 1991,'The Cataract Explosion: The

Changing Pattem of Diagnosis of Patients Attsnding an Ophthalmic Outpatient

Department', Elê, 5: P369-372.

252 References

Belkin M., Jacobs D.R., Jackson S.M. et al. 1982'senile catafacts and myopia'' Ann

Ophthalmol, 1 4: P49-50.

Berndt R.M., Berndt C.H.1977, The Wortd of the lirsl Australian, Ure Smith, Sydney'

Bhat K.S. 1987,'Nutritíonal status of thiamic, riboflavin and pyridoxine in cataract pat¡ents',

Nutr ReP lnt,36: P683-692.

Bhatnagar R., West K.P., Vitale S., Sommer A., Joshi S., Venhataswamy G. 1991, 'Risk of

Cataract and History of Severe Diarrheal Disease in Southern lndia', Arch Ophthalmol

May, 109: P696-699.

,Etat Bietti G.B. 1g57, actuel De La Chimiotherapie Et De L'Antibiotherapie Du Trachome" Bull

World Health Organ,16: P975-994.

Black E.C., Cleland J.B. 1938,'Palhological lesions in Australian Aborigines in Central

Australia (Granites) and Flinders Range', J Trop Med Hyg,41: p69.

Blomdahl S., Norell S. 1984, 'Perforating eye injury in the Stockholm population. An

epidemiological study', Acta Ophthalmol (Copenh), June,62 (3): p378-90.

Bochow T.W., West S.K., Azar 4., Munoz 8., Sommer A., Taylor H. 1989, 'Ultraviolet Light

Exposure and Risk of Posterior Subcapsular Cataracts', Arch OPhthalmoLl0T: p369-

372.

Boldt J. 1904, Trachoma, Hodden and Stoughton, London.

Brennan R., Patel M., Hope A. 1989, 'Gonococcalconjunctivitis in Central Australia', (lettar)

Med J Aust, 150: P48-49.

Brian G, Danzell J, Nangala S, Hotlows F.C. 1990, 'Basic ophthalmic assessment and core

workshops for rural health workers', Aust and NZ J Ophthalmol, Feb,1 I (1 ).

253 References

Brilliant L.8., Grasset N., Pokhrel R.P., Lepkowski J.M', Brillant G.E., Hawks W.N.,

pararajasegaram R. 1983 a, 'Associations among cataract prevlaence, sunlight hours

and altitude in the Himalayas', Am J Epidemiol,118 (2): p250'2ü-

Brilliant L.8., Lepkowski J.M., Musch D.C. 1983 b, 'Reliability of ophthalmic diagnosis in an

epidemiologic survey', Am J Epidemiol,118 (2): p265-279.

Brilliant GE, Brilliant LB. 1985, 'Using social epidemlology to underctand who stays blind and

who gets operated for cataract in a rural setting', Soc Sci Med,21 (5): p553-558.

Brill¡ant LB, Pokhrel RP, Grasset NC, LepkowskiJM, Kolstad A, Hawks WN, Pararajasegaram

R, Brilliant GE, Gilbert S, Shreslha R, Kuo J. 1985, 'Epidemiology of blindness in

Nepal', BullWorld Health Organ,63 (2): p375-386.

Brunham R.C., Laga M., Simonsen J.N., Cameron D.W., Peeling R., McDowellJ., Pamba H.,

Ndinya-Achola J.O., Maitha G., Plummer F.A. 1990, 'The Prevalence ol Chlamydia

trachomatis lnfection Among Mothers of Children with Trachoma', Am J Epidemiol,

132 (5): P946-952.

Bucher P.J.M., ljsselmuiclen C.B. 1988, 'Prevalence and causes of blindness in the Northem

Transvaal', Br J Ophthalmol, 72: p721 -726.

Carmichael T.R., Gibson 1.H., Kustnan H.G. 1982, 'Blinding trachoma: a public health

challenge', S Afr Med J,61 (1): P5€.

Cedrone C., Gatli M.G., Cerull¡ L. 1987, 'Public Health Approach in Trachoma ControlAcvitity',

Rev-t nt-Trach - Pathol-Ocul-Trop'Subtrop-Sante'Publiqu e, 8t: pi 03- 1 4'

Chatterjee A., Milton R.C., Thyle S. 1982,'Cataract prevalence and etiology in Puniab', BrJ

Ophthalmol, 66: P35-42.

Chumbley L.C., Thompson LM. 1988, 'Epidemiology of Trachoma in the West Bank and Gaza

Strip', Eye, 2 (Pt5): p463-470.

254 References

Chumbley L.C., Viswalingam N.D., Thomson LM., Zeidan M.A. 1988, 'Treatment of trachoma

in the West Bank', Eye,2 (Pt5): p471-475.

Clayton R.M., Cuthbert J., Seth J., Phillips C.1., Bartholomew R.S., McK Reid J' 1984'

,Epidemiological and other studies in the assessment of factors contributing to

cataraclogensis', Cröa Foundation Symp, 106: p25-47.

p125' Cleland J.B. 1928, 'Diseases Amongst the Austral¡an Abodgines', J Trop Med Hyg,31:

cochrane w.G., Gemmelw. 1977, Sampling Techniques, wiley, New York.

Collier L.H. 1974, Rev tnt Trachme,(S1),7: quoted in Jones 1975.

Collman G.W., Shore D.L., Shy C.M., Checkoway H., Luria A.S. 1988, 'Sunlight and other dsk

factors for cataracts: An epidemiological study', Am J Public Health, Nov;78 (11):

p1 459-1 462.

Condon R. 1991,'Gonococcal Conjunctivitis in the Ngaanyat¡arra Homelands of Western

Australia April-June 1991. lnvestigat¡on and outbreak control measures', Occasional

Paper 44, Health,W.A..

Cook J. ,Account ol the Voyage Round the Wortd in the Years 1768 1769 1770 1771 1773

London iii, P634.

Cooper R.L., Coid D., Constable l.J. 1986, 'Trachoma 1985 updats in W.A.', Aust NZ J

Ophthalmol, Nov, 14 (4): P319-323.

Courtright P., Sheppard J., Schachter J., Said M.E., Dawson C.R. 1989, 'Trachoma and

blindness in the Nile Delta: current patterns and proiections for the future in tho rural

Egyptian population', Br J Ophthalmol,T3: p536-540.

Courtright P., Sheppard J., Lane S., Sedak 4., SchachterJ., Dawson C.R. 1991,'Latrine

ownership as a protective factor in inflammatory trachoma in Egypt', Br J Ophthalmol,

75: p322-325.

255 References

Crotty J.M., Mann 1., McLean D.M. 1959, 'Trachoma in Northern Australia. Bac{eriologic and

virologic aspects', Am J Ophthalmol,4T: p503'

Cutts F.T. 1988, 'The use of the WHO cluster survey method for evaluating the ¡mpact of the

expanded programme on immunization on target disease incidence', J Trop Med Hyg,

91: p231-239.

DampierW.M. 1688, ANewVoyageAroundtheWorld,ChapterXVl :quotedinCleland 1928.

Dana M.R., Telsch J.M., Enger C., Joyce E., Santoli J.M., Taylor H'R. 1990, 'Msual lnpairment

in a RuralAppalachion Community Prevalence and Causes', JAMA, Nov 14 ;26,4 (18):

p2400-2408.

Darougar S., Jones 8.R., Viswatingam N., Poirier R.H., Allami J., Houshmend 4.,

Farahmandiam M.4., Gibson J.A. 1980, 'Family-based suppressive intermittent

therapy ol hyperendemic trachoma with topical oxyletracycline or oral doxycycline',

Brit J Ophthalmol,64: P291-295.

Darougar S. and Jones B.R. 1983, 'Trachoma', Br Med Bull,39:p117-122-

Dawson C.R. and Schachter J. 1967, 'Trachoma in Jamaica. Epidemiologic and microbiologic

observations on mild disease', Am J Ophthalmol,63: pl408.

Dawson C.R., Daghfous T., Messadi M., Hoshiwara 1., Schacter J. 1976, 'Severe endemic

trachoma in Tunisia', Brit J Ophthal,60: p245.

Dawson C.R., Jones 8.R., Tarizzo M.L. 1981a, 'Field guide to trachoma control', World Health

O rg ani sation, Ge neva, P38-47.

Dawson C.R., Daughfous T., Whitcher J., Messadi M., Hoshiwana T., Trik¡ F., Chalgrah F.,

Briones O., Yoneda, Schachler J. 1981b, 'lnterm¡tlent trachoma chemotherapy: A

controlled trial of topical tetracycline or erythromycin', Bull World Health Organ, 59:

p91-97.

256 References

Dawson C.R., Schwab l.R. 1981, Epidemiology of cataract a major cause of preventable

blindness, Butt World Health Organ,59: p493-501.

Dawson C.R., Daghfous T., Hoshiwara 1., Ramdhane K., Kornoun M., Yoneda C', Schachter J' ,Trachoma 19g2, therapy with topicaltetracycline and oral erythromycia; a comparative

lnal', Bult World Health Organ,60 (3): p347-355.

Dawson C.R. and Schachter J. 1985, 'strategies for treatment and control of blinding

trachoma; cost effectiveness of topical or systemic antibiotics', Rev lnfecl Drs, Nov-

Dec,7 (6): P768-773'

Dethlefs R. 1g82, 'The Trachoma status and blindness rates ol selected areas of Papua New

Guinea in 1979'80', Aust J Ophthalmol, Feb,10 (1): pl3-18.

Dobson A.J. 1g87, 'Trends in Cardiovascular Risk Factors in Australia, 1966-1983: Eviderrce

from Prevalence Surveys', CommunityHealth Stud,11' (1): p2-14.

Douglas F., Elgar W., Sharrock D., Powers J. 1987, (20), 'Chlamydia infection - Longitudinal

study of Mothers and lrìfants', Menzies School ol Health Research Annual Report

June 1 986-Junel 987, (20): P94-95.

Douglas F., Gardner 1., Foreman A., Asche V., Shanock O., Powers J., Mathews J. 1987, (19)'

,Chlamydia infection and chronic otitis media', Menzies School ol Health Research

Annual Beport Junel 986-Junel 987, (19): p91-92.

Drummond M.F., Yates J.M. 1991,'Clearing the Cunent Backlog in a (Not So) Developing

Country', EYe, 5: P481-'186.

Duke-Etder S. 1965, System of Ophthalmotogy. Diseases of the outer eye;Vol Vlll London,

Henry Kimpton,S (1): P260.

Dunn F.L. i985, 'Socio medical contributors to trachoma research and intervenÅon', Review

lnfect Dis, Nov-Dec: P783-786.

257 Ref erences

Easterman 4., Mac Harper T., Rohrsheim R., Roder D. 1988, South Auslralian Health SfaÛsfi'cs

Chart Book, Epidemiology Branch, South Australian Heatth Commission, Adelaide.

Ederer F. 1g73, 'Shall we count number of subjects or number of eyes?', Edìtorial Arch

Ophthalmol, 39: P1-2.

Ederer F. 1g75, 'Palíent bias, investigator bias and the double-masked procedure in clinical

trails', Am J Med,58: P295-299

Ederer F., Hiller R., Taylor H. 1981,'senile lens changes and diabetes in two population

studies', Am J Ophthalmol, gl: p381-395.

Ederer F. 198g, 'Methodologicat problems in eye disease epidemiology', Epidemiol Rev,5:

p51-66.

Editorial. 1972,'Trachoma and related disorders', Med J Ausf' 2: p859'

Edwards B. 1990,'Putuna Kuliplpai: lCannot Understand', New Doà'tor,53: p10-13.

Edwards F.M, Wise P.H, Craig R.J, Thomas D.W, Murchland J.B. 1976, 'Visual acuity and

retinal changes in S.A. Aborigines', Aust. NZ J Med 1976,6: p205-209'

Elliot R.H. '1920, Tropicat Ophthalmotogy, Qrtord University Press, London.

Elphinstone J.J. 1971,'The health of Aust. Aborigines with no previous association with

Europeans' , Med J Aust,2: P293.

Eyre E.J. (1854), Journals ol Expeditions of Discovery into Central Australia, Australian

Facsimle Edition No. 7 (1964) Adelaide: Libraries Board of S.A.'

Faal H., Minassian D., Sowa S., Foster A. 1989, 'National survey of blindness and low vision in

Gambia: results', Br J Ophthalmol, Feb,73 (21l. Ñ2-87

Feibet R.M. 1983, 'John Vetch and the Egyptian Ophthalmia' Suru Ophthalmol,2S:p128'

258 References

Ferris F.L, Ederer F. 1979, 'Exlernal monitoring in multi clinical trials. Applications from

ophthalmo logic studi es', C ti n Ph a r m aco I Th e r, 25 : p7 20 -723

Fleiss J.L. 1981 , Statistical methods for rates and proportions, 2nd ed. New York, John Wiley

and Sons.

Fletcher L.J., Gordon R.C. 1990, 'Perinatal Transmission of Bac{erial Sexually Transmitted

Diseases Part ll', J Fam Prac,30, (6): p689-696.

Flynn F. 1g57, 'Trachoma arnong natives of the Northern Territory of Australia', Med J Aust,2:

p269.

Forsey T., Darougar S. 1981, 'Transmission ol chlamydiae by the housefly', Brit J Ophthalmol,

65: p147-150.

Foster A. 1987, 'Cataracl Blindness in Africa', Ophthalmic Surg, May;l8 (5): p384-388.

Freeman P., Holmes W., Tozillo P. 1985, Nganampa Health Council - medical reprt 1984,

Nganampa Health Council, Alice Spings, N.T..

Friedrech R. i982, 'Eye disease in the Navajo lndians', Ann Ophthalmol, Jan,14 (1): p38-40.

Gibson R.A., Sanderson H.F. 1980, 'Observer variation in ophthalmololgy', Br J Ophthalmol,

64: p457-460.

Goldstein H. 1980,'The reported demography and causes of blindness throughout the

world', Adv OPhthalmol,40: P1-99.

Goodrruin P. 1g83, 'Training health workers: what needs to be taught and who should teach it',

Soc Sci Med, 17'. Pl819-1825'

Graham D.M., Nichols R.L., Mann l. 1973, Med J Aust Aug 25: p353-360.

Grayston J.F., Wang S., Yeh L., Kuo C. 1985, 'lmportance of reinfection in the pathogenesis

of trachoma', Rev lnfect Dis, Nov-Dec: p717-725'

259 References

Grey G.H. and Cleland J.B. 1933, 'Some Pathological Conditions in Central Australian

Aborigines', J TroP Med HYg, xxxvil'

Grin T.R., Nelson L.B. Jeffers J.B. 1987,'Eye injuries in childhood', Paediatrics, Jul,80 (1):

p13-17.

Guerra p. 1957,'Racial and regional differences in the epidemiology and clinical

manifestations of Trachoma" Bull world Health organ,16: p995-101 1.

Hamilton J.B. 1950, 'A New Survey of Tasmanian Blindness 1948', Trans Ophthalmol Soc

Aust (BMA), 10: P42.

Harding J.J. 1982, 'Calaract: sanitation or sunglasses?', LanceÍ, 1: p39.

Harding J.J., Rixon K. 1981, 'ls diarrhoea a major cause of cataract in some tropical countries?',

Metab Pediatr Syst Ophthalmol, 5: p161-166.

Harding J.J., von Heyningen R. 1987,'Epidemiology and dskfactorsforcataract', Eye,1(51:.

p537-541.

Hardy D., Surman P.G., Howarth W.H. 1967, 'The cytology on coniunctival smears from

Aboriginal school children at Yalata, South Australia, after improved hygiene

conditions and treatment with oxytetracycline and systematic sulphonmetoxinê', AmJ

Ophthalmol, 63: P1 538-1 540.

Hennekens C.H., Buring J.E. 1987, Epidemiology in Medicine, Little Brown and Company

Boston.

Hiller R., Kahn H. 1976, 'senile cataract extraction and diabetes', Br J OphthalmoL60: p283-6.

Hiller R., Giacommet L., Yuen K. 1977, 'sunlight and cataract: An epidemiologic investigation',

Am J Epidemiol, 105: P450-459.

260 References

Hiller R., Sperduto R., Ederer F. 1983, 'Epidemiologic assciations with cataract ¡n the 1971-

1972 National Health and Nutrition Examination Survey', Am J Epidemiol,118: p239-

249.

Hiller R, Sperduto R, Ederer F. 1986, 'Epidemiologic associations with nuclear, cortical and

post subcapsularcataracts', Arn J Epidemiol,Dec:124 (6): p916-925'

Hollows F.C., Moran D. 1981, 'Cataract the Ultraviotet Risk Factor', Lancet, Dec 5: p1249-1250

Hollows F.C. 1982, 'A view of some Aboriginalinlections', New Doc'tor,26:p21-26.

Hollows F.C. 1g85, 'Communily based action for the control of trachoma', Review of lnfecl Dis,

Nov-Dec: P777-782.

Hollows F.C. 1989,'Trachoma"DowntheTrack"', MedJ Aust,Aug 21,151:p182-183.

Houston S., Legge D. 1992, 'Aboriginal health research and the National Aboriginal Health

Strategy, Aust J of Public Health,16, 2: p114-115.

Hyman L. 1987, 'Epidemiology of eye disease in the elderly', El€,1 (pt2): p330-341.

lslar M., Chirambo M., Belkin M. 1982, 'Ocular inþdes in Malari', BrJ Ophthalmol,6:p145.

Jones B.R, Collier LH, Smith CH. 1959, 'lsolation of virus inclusion blennorrhoea, Lancet,l;

P902.

Jones B.R. 1g75, 'The prevention of blindness from trachoma', Trans Ophthalmol Soc UK,

95: P16.

Jones B.R. 1978, 'symposium on prevention of blindness', Irans Ophthalmol Soc UK,98;

p282-86.

Jones B.R. 1gBO, 'Changing concepts of trachoma and its control', Tnns OphthalmolSoc UK,

100: P25-29.

261 Ref erences

Dawber T'R' 1975' Kahn H.A, Leibowitz H.M., Ganley J.P., Kini M., Colton T., Nickerson R',

,standardizing diagnostic procedures" Am J Ophthalmol,24: p366-390

Dawber T'R' Kahn H.A, Leibowitz H.M., Ganley J.P., Kini M.M', Colton T., Nickerson R'S',

1977, 'The Framingham eye study. Outline of major prevalence findings', Ám J

EPidemiol, 106: Pl7'

,The Med Aust, Kamien M. 1976 a, physical health of Aboriginal Children in Bourke NSW', J

Special Supplement on Aboriginal Health,6, 1, 3: p33-37'

Med Aust, Kamien M. 1976 b, 'The physical health of Aboriginal Adults in Bourke NSW', J

Speciat Supplement on Aboriginal Health,6, 1, 3: p38- 44'

Katz J., Zeger S.L., Telsch J.M. 1988, 'Village and Household Clustering of Xerophthalmia

and Trachoma', lnt J Epidemiol,lT (4\: p865-869'

Kelsey J.L., 19g6, Methods in Observational Epidemiology, in Monographs in Epidemiology

and Biostafistics, eds. Thompson W.D. and Evans 4.S., Oxford University Press,

Oxford: Vol. 10, ChaP' 8'

ol Kinoshita J.H., Varna S.D., Fukui H.N. 1979, 'Aldose reductase in diabetic complications

the eye', Metabol¡sm, 28: P462-469'

,The Kok p.W. 19g3, Epidemiology of Trachoma Blindness in Southem Aîrica', Soc Sci Med,

17 (22): P1709-1713.

Kupka K., Nizetic 8., Reinhards J. 1968, 'sampling studies on the epidemiology and control

of trachoma in southern Morocco" Bult world Health organ,39: p547-66'

Leibowitz H.M., Kruegen D.8., Maunden L.R et al. 1980, 'The Framingham Eye Study

Monograph', Suru Ophthalmol, 24: p366-61 0.

Lerman s. 1980, Radiant energy and the eye, New York Macmillan, p31.

262 References

Leske M.C. and Sperduto R.D. 1983,'The Epidemiology of senile cataracts: A review', ,4mJ

EPidemiot, Aug, 118 (2): P152-165'

Leske M.C., Cynlack L.T., Wu S.Y. 1991,'The Lens Opacity Case-Control Study', Arch

Ophthalmol, 109: P224-251'

Lieban R. 1g78, Sex differences and cultural dimensions of medical phenomenom in a

philippines setting. ln Culture and Curing, eds. Morley P. and Walter R., Allen,

London.

Loewenthal R., Pe'er J. 1990, 'A prevalence survey of Ophthalmic diseases among the

Tud

MacCallan A.F. 191g, Trachoma and its cornptications in Egypt, Cambddge University Press.

p369-41 Macçallan A.F. 1931 , 'The epidemiology of trachoma', Br J OphthalnþL 15: I '

MacDonald A.F. 1965, 'Causes of blindness in Canada', Can Med Assoc J,92:p2ú'279.

Maddocks 1., Maddocks D. 1991,'Control: a central focus of health', PHA Conference

Abslracts, Aust J of Public Health,15' 4: p329'

Maddocks l. 1gg1,'Ethics in Aboriginal research: a model lor m¡norities orlor all?', PHA

Conference Abstracts, Aust J of Public Health,15, 4: p329

Majouk J.F. 1966, 'A study of trachoma and associated infection in the Sudan', Bull World

He alth Org, 35'. P262-272.

Malaty R., Zaki S., Said M.E., Vastine D.W., Dawson C.R., Schachter J. 1983, 'Extraocular

infeclions in children in areas w¡th endemic trachoma', The J of lnfec Drs, June, 1(l

(6): p853.

Mann l. 1954a, Ophthatmic suvey of the Kimberley Division of W.A., Public Health

Department of W.A. Publication'

263 Ref erences

Mann l. 1g54b, Ophthalmic suNey ol the Eastern Goldfields Division of W.A., Public Heatth

Department of W.A. Publication.

Mann l. 1g57a,. 'Report on Ophthalmic findings in Warburton Range Natives of Central

Australia', Med J Aust, 2: P610.

Mann l. 19S7b, 'Probable Origins of Trachoma in Australasia', BullWorld Health Organ,16:.

p1165-1187.

Mann l. 1967, 'Correlation of race and way of life in Australia and the terdtory of Papua New

Guinea with incidence and severity of clinical trachoma', Am J Ophthalmol, May: p

13021276-1309/283.

Mann l. and Rountree P. 1968, 'Geographic ophthalmology: A report on a recent survey of

Australian Aboriginals', Am J Ophthalmol,66: p1020.

Marshall C.L. 1968, 'The relationship between trachoma and piped water in a developing

area', Arch Environ Health, 17: p215-220-

Martinez G.S., Cambell A.J., Reinken J., Allan B.C. 1982, 'Prevalence ol ocular disease in a

population study of subjects 65 years old and olde/, Am J Ophthalmolmol, Aug, 94

(2): p181-189.

Manx R. 1989, 'social factors and trachoma: A review of the literature', Soc Sci Med,29 (11:.

p23-34.

Matthews J. 1987, 'A Strategy to Prevent Chlamydial Disease', Menzies School of Health

Research Annual Repoft, June 86-June 87; (18): p89-91.

Matthews J., Asche V., Douglas F., Peach H., Woods M. 1987, 'Prospects for the Elimination

Chlamydial Disease', Menzies Schoolof Health Research Annual Repoft, June 8&

June, : P78-79.

264 Ref erences

Mathur G.M., Sharma R. 1970, 'Prevalence of trachoma and other comrnon eye diseases' lnd

J Med Fes,58: P1085.

Mburu F.M., Steinkuller P.G. 1983, 'Ocular problems, resource needs and health policy', Soc

Sci Med, 17: P1683-1691 .

McGuiness R., Hollows F.C., Tibbs J., CampbellD.1972, 'Labrador keratopathy in Australia',

Med J Aust, 2: P1249-50.

Meredith S.J., Peach H.G., Devanesan D. 1989, 'Trachoma in the Northem Territory of

Australia', Med J Ausf, Aug 21 : Pl51 .

Met¡er C.C. 1947, History of Medicine, Ch1 4 Philadelphia: The Blakiston Co., : pl005-1022'

Miller R.J., Fujino J., Nefzger M.D. 1967, 'Lens findings in atomic bomb survivors. A review ol

major ophthalmic surveys at the Atomic Bomb Casualty Commission 1949 - 1962',

Arch Ophthalmol, 78: P697 -704.

Milne J.S., Wlliamson J.1972, 'Msual acuity in older people', Geront Clin,14:. p249-56.

M.l.M.S. Annual 1990, lnternationalMedicalStatistics, Crows Nest, Sydney, Australia.

Minassian D.C. 1988, 'Epidemiologicalmethods in prevent¡on of blindness', Eye,2: p3-12.

Minassian D.C., Mehra V., Verrey J.D. 1989, 'Dehydrational crises: a major risk factor in

blinding cataract', Br J Ophthalmol, T3: p100-f 05.

M.M.W.R. 1982, 'Prevention of blindness: Trachoma conlrol', Ø;33 (41): p555-6,p561.

Mobbs R. 1986, 'But I do Care! Communication Ditficullies affecting the quality of caie

delivered to Aborigines, Med J Aust (Special Supplement on Aboriginal Health), Junø

23,144.

265 References

Mohon M., Sperduto R.D., Angra S.H.,Milton R.C., Mathur R.L., Undenvood B.A', Jatfery N',

Pandya 8., Chhabra V.K., Vaipaayee R.8., Kalra V.K., Sharma Y.R. 1989, 'lndia - US

case control study of age-related cataracts', Arch Ophthalmol,l0T: p670-676.

Monnickendam M.A., Darougar S., Treharne J.D., T¡lbury A.B. 1980a, 'Guinea-pig inclusion

conjunctivitis as a modelfor the study of trachoma: clinical, microbiological, serological

and cytological studies of primary infection', Br J Ophthalmol, il:p279-283'

Monnickendam M.A., Darougar S., Treharne J.D., Tlbury A.B. 1980b, 'Development of

chronic conjunctivitis with scaring and pannus, resembling trachoma in guinea-pigs',

Br J OPhthalmol,64: P284-290.

Monnickendam M.A., Pearce J.H. 1983, 'lmmune Responses and Chlamydial Infections', Br

Med Bull,39: P187-193.

Moore M.C., Howarth W.H., Wilson K.J., Derrington 4.W., Surman P.G. 1965, 'Clinical and

laboratory assessmenls of trachoma in SA', Med J Aust,2'. Ø41 .

Morsey S. 1gBO, 'Health and illness as symbols of socialditferentiation in an Egyptian village',

AnthroP Quart, P1 53-1 61.

Mulvey G., Endean C. 1991,'Ernabella noncommunicable diseases survey', AustJ Public

Health,15,4: P349.

Nasr A.M. 1991,'Eyelid complication in trachoma diagnosis and management', ACTA

Op hth al mo lo gi ca, 69 : P200-204.

Nathan P., Japangka D.L. 1983, 'Two-way Medicine: A Misnomer', Health Business,

Heineman Educational, Australia, p68-92.

National Aboriginal Heatth Strategy Working Party, 1989, The Report: A National Aboriginal

Heafth Strategy., AGPS Canberra.

266 Ref erences

National Health and Medical Research Council 1992, Guidelines on ethical matters in

Abriginat and Torres Strait tstander heafth research, N.H.M.R.C',Canberra.

Nganampa Health Council 1987, 'An Environmental and Pubtic Health Review within the

Anangu P1jantiatiara Lands', Report of Uwankara Palyanyku Kanyintiaku, Dec.

Olurin O. 1971, 'Causes of Enuculation in Nigeria', Am J Ophthalmol, T6: p987-991'

,The Packer A.D. 1961, Health ol the AuStralian Native" oceania,32: p60.

Paisley J.W., Laver 8.A., Mc lntosh K., Globe M.P., Schachter J., Rumach C. 1984,

,pathogens associated with acute lower respiratory tracl infection in young children',

Pediatr lnfect Dis,3: P15-19.

PeaCh H., Piper P., Devanesan D., DixOn 8., JetferieS C., Braun P., NelSOn D., Kruger G.,

Boulder H. 1987, 'Trial of antibiotc eye drops for the prsvent¡on of trachoma in school-

age Aboriginal children', Menzies School of Health Research Annual Report June

1986-87, (18): P 7a.

pirie A. 1972, 'Photo-oxidation of proteins and comparison ol photooxidation proteins in

those of the cataracÍous human lens', /srael J Med Sa,8: p1567-73.

pisa Z. 1989, 'Multinational Monitoring of Trends and Determinants in CardiovasqJlar Disease.

lnternalional Comparisons: An Overview', lnt J Epidemiol 18, 3: S19.

pizzarello L.D. 1987, 'The dimensions of the problem of eye disease arnong the elderly',

Ophthalmology, Sept, 94 (9): pt 191-1 195.

pizzarello L.D. 1990, 'Training Health Care Workers in Ophlhalmic Care: The Experience of

Helen Keller lnternalional', lnt Ophthal Clinics, S0' (1).

potter A.R. tgg1, 'Causes of blindness and visual handicap ¡n the Csntral Afdcan Republic',

Br J Ophthalmol, TS: P326-328.

267 References preece p.M., Anderson J.M., Thompson R.G. 1989, 'Chlamydia trachomatis infection in

infants: a prospective study', Arch Dis Child, &: p525-529'

prost A., Negrel A.D. 1989, 'Water, Trachoma and coniunctivitis', BullWorld Health Organ,67

(1): p9-18.

Rapoza P.A, Quinn T.C., Kiessling R., Gree R., Taylor H. 1986, 'Assessment of neonatal

conjunctivitis with a direct immunofluorescent monoclonial antibody stain for

chlamydia', JAMA, 255: P3369-3373.

Reacher M.H., Huber M.J.E., Canagaratnam R., Alghassany A', 1990, 'A trial of surgery for

trichiasis of the upper lid f rom trachoma', Brit J of Ophthalmol, T4: pl 09-1 1 3.

Redmond K.B. 1949, 'The incidence and causes of blind eyes with some remarks on their

prevention', Trans Ophthalmol Soc Ausf,9: p192.

Reinhards J. 1967, 'Trachoma control in the European region', World Health Organ Chron,21:

P57-61.

Reinhards J., Weber 4., Nizetic 8., Kupka K., Maxwell-Lyons P' 1968, 'Studies ¡n tho

Epidemiology and Control of Seasonal Conjunctivitis and Trachoma in Southern

Morocco', Bult World Heafth Organ,39: p497-545.

Rhodes J., Asche V., Schullen M., Gardner l. 1987, 'Are dogs a reservor lor Chlamydia

Trachomatis infections?', Menzies School of Health Research Annual Reprt 198&

1987:: p80-81.

Roper-Hall M.J. 1978, 'Prevention of blindness lrom trauma', Trans Ophthal Soc UL,98;

p313-315.

Rosenthal A.R. 1980, 'The Framingham Eye Study (Editorial)', Suru Ophthalmol, May-Jun,

24(6): p611-613.

268 Ref erences

Rosenthal A.R. 1g88, 'Methods of epidemiological investigation: discussion paper', J R Soc

Med, SeP, 81 (9): P530-541.

Rosner B. 1982,'statistical methods in ophthalmology: an adiustment for the interclass

correlation between eyes', Biometics, 38: p105-1 14.

Royal Australian College of Ophthalmologists 1980, Il¡e NationalTrachoma and Eye Health

Program: RePort, SYdneY, R'A.C'O..

Said M.E., Goldstein H., Korra 4., El-Kashlan K. 1970, 'Prevalence and causes of blindness in

urban and rural areas of Egypt', Public Health Rep,85: p587-99.

Said M.E., Goldstein H., Korra 4., El-Kashlan K. 1973, 'Blindness prevelence rates in Egypt',

Health Servs Reg,88: P89-96

Salonen J.T., Kottke T.E., Jacobs D.R., Hannan P.J. 1986, 'Analysis of community-based

cardiovascular disease preventio studies - evaluation issues in the North Karelia

Project and the Minnesota Heart Health Program', Int J Epidemiol,lS:p176-182'

Sandford-Smith J. 1984, 'Ophthalmology in developing countries' , Br Med J (Clin Res), Sept

29, 289 (6448): Pe1 1 -81 3.

Schachter J., Luin L., Gooding C.A.et al. 1975,'Pneumon¡tis following inclusion

blennorrhea', J Pediatr, 87 : p779'780.

Schachter J. 1978, 'Chlamydia infeclions', N Eng J Med,298: p428-435, 490-495, 540-549.

Schachter J., Dawson C.R. 1981, 'Chlamyial infections, a world wide problem: epidemiology

and implications for trachoma therapy', Sex Transm Drs, 8: p167'74.

Schachter J. 1985, 'Overyiew ol Chlamydia trachomatis lnlection and the Réquirements for a

Vaccine', Rev lnf Dis,7 (6): P713-715

269 References

Schachter J., Grossman M., Sweet R.L., Holt J., Jordon C., Bishop E. 1986 a , 'Prospeclive

study of perinatal transmission ol C. trachomatis', JAMA,255: p3374-3377.

Schachter J., Sweet R.L., Grossman M., Landens D., Robbie M., Bishop E.1986 b,

'Experience with the routine use of erythromycin lor chlamydia infection in

pregnancy', N Engl J Med,314: p276-279.

Schaetfer R.L., Mendenhall W., Otl L. 1990,'Elementary Survey Sampling', Boslon, PWS-

Kent Pub. Co..

Schebeck B. 1986,'After successful lield work: what to do with all the'material"?, Ausf

Aboriginal Studies, 1 : P52-58.

Schein O.D., Hibberd P.L., Shingleton 8.J., Kunzweiler T., Frambach D.A', Seddon J'M.,

Fontan N.L., Vinger P.F. 1988, 'The spectrum and burden of ocular iniury',

Ophthalmology, Mar, 95 (3): P300-5.

Schneider M. 1946, 'A sociological study of the Aborigines in the N.T. and their eye disease',

Med J Aust, 1: p99.

Schwab L. 1990 a, 'Cataract Blindness in Developing Nations', lnt Ophthalmol Clinics, Winter,

30 (1): p6-18.

Sclrwab L. 1gg0 b, 'Blindness from trauma in developing nations', lnt Ophthalmol Clinics,30

(1): p28-2e.

Seddon J.M., Christen W.G., Manson J.E., Buring J.E., Sperduto R.D., Hennekens C.H'

1g91, 'Low-Dose Aspirin and risk of Cataracts in a Randomized Tdal of US Physicians',

Arch Ophthalmol, Feb, 109: P252-258.

Sheffield V.M. 1983, 'Training for primary and preventive eye care', Soc Sci Med, 17:. p1797-

808.

Shingleton B.J. 1991 , 'Eye lnjuries', N Eng J Med. ,325,6: p 408-413.

270 References

Smith R.J.H. 1990, 'Editorial', Br J Ophthalmol,T4:p324.

Smith R.J.H. 1991, 'Trachoma still undefeated', Editorial, Br J Ophthalmol, T5: p321.

Social Security Department 1991, 'Guides to the Administration of the Social Security Act',

Ref 10.801 , socialseatrity Depanment, Adelaide, south Australia. sole G.D. 1gg7,'lmpact of cattte on the prevalence and severity ol trachoma" BrJ

Ophthalmol, 7'l : P 873-876.

Sommer A. '1977, 'Cataracts as an epidemiologic ploblem', Am J Ophthalmol, 33: p334-339.

Sommer A. 1986, 'Attack Blindness', Am J Ophthalmolmol, Sep 15, 102 (3): p387-389.

Sommer A. 1988, 'Avoidable Blindess', Aust NZ J Ophthalmol, Feb, 16 (1): p31-35'

Sorsby A. 1966, 'The incidences and causes of blindness in England and Wales 1948-1962.

Reports on Pubtic Health and Medical Subiecls, (London)' 14'

Stamler J. 1989, 'Oppurtunities and Pitfalls ¡n lnternational Comparisons Related to Pattems,

Trends and Determinants ol CHD Mortality', lnt J Epidemiol, (supp),18' (3): p s3-s18.

Steinhuller P.G. 1983, 'Cataracl: The leading cause of blindness and vision loss in Africa', Soc

Sci Med,17 (221: P1693'1702.

Stirling E.C. 1894, lntercol Quart, Journ of Med and Surg vol i,(3): p218, quoted in Cleland

1 928.

Sturt C. (1883). 'Two Expeditions into the lnterior of Southern Australia, 2 vols. London',

Facsimile reP, Adelaide (1963).

Sutter E.E. and Ballard R.C. 1988, 'Community participation in the control of Trachoma in

Gazankula', Soc Sci Med,17: Pl813'37

271 Ref erences

,social Syme 19g6, Determinants of Heatth and Disease'.ln Public Health and Human

Ecology, ed Last J.M., Appelton-Century-Crofts, Norwalk, Conneclicut, p953-970.

Tabbara K.F and Ross-Degnar D. 1986, 'Blindness in SaudiArabia', JAMA, Jun27,255 (24):

p3378-3384.

Tabbara K.F., Taylor P.B. 1988, 'Trachoma anþng schoolchildren in Al-Ahsa', Saudi Med J,9

(1): p5a-58.

Tabbara K.F. 1990, 'Trachoma: Have we advanced in the last 20 years?', lnt Ophthal Clinics,

Winter,30 (1): P23-27.

Tang, Chong, Huong, Wang '1957, China Med J,75:p429, quoted in 'Editorial' Med J Aust

'1972.

Taylor H.R. 1977, 'Blindness in Australian Abotiginals', Aust- J Ophthalmol 5: p155.

Taylor H.R. 1g8O a,'Prevalence and causes of blindness in Australian Aborigines', Med J

Aust, Jan 26: P71-76.

Taylor H.R. 1980 b, 'The environment and the lens', BrJ Ophthalmol, ô4: p303-310.

Taylor H.R., Velasco F., Sommer A. 1985, 'The Ecology of Trachoma: an epidemiological

study in Southern Mexico', Bull World Health Organ,63 (3): p559-567.

Taylor H.R., Sommer A. 1985, 'Risk factor studies as an epidemiologic lool', Review of lnfect

Dis, Nov-Dec:765-767.

Taylor H.R. 1987a,.'strategies for the control of lrachoma', Aust NZ J Ophthalmol, May, 15 (2):

p139-43.

Taylor H.R. 1987 b, 'Trachoma Research: Laboratory and Epidemiologic Aspects', Rev'lnt-

Trach- P athol-Ocul-TroVSubtrop-Sante' Publiqu e, ü: p25-28.

272 Ref erences

Taylor H.R. 1987 c, 'Trachoma Grading: A New Grading Scheme', Rev-lnt-Trach-Pathol-Ocul-

Trop-Subtrop-Sante-Publique, 64: p175-181.

Taylor H.R., Wesl S.K., Katala S., Foster A. 1987, 'Trachoma: evaluation of a new grading

scheme in the United Republic of Tanzania', Bull World Health Organ,65, (4): p485-

488.

Taylor H.R. 1988, 'A simple method for assessment of association between synanlhropic flies

and trachoma;, Am J Trop Med Hyg,38 (3): p623-627.

Taylor H.R, Wsst S.K., Rosenthal 4.R., Munoz 8., Newland H.S., Abbey H., Emmett E.A.

1988, 'Eflect of ultraviolet rad¡at¡on on cataract formation', N Engl J Med,319; pl429-

1433.

Taylor H.R., West S., Mmbaga B.B.O., Katula S.J., Turner V., Lynch M., Munoz 8., Rapoza P.

1989, 'Hygiene Factors and lncreased Risk of Trachoma in CentralTanzania', Arch

Ophthalmol, Dec, 107: P1821-1825.

Tedesco L.R. 1980, 'Trachoma and environment in the Northern Territory of Australia', Soc

Sci Med,14: P111-117.

The ltalian-Ameñcan Cataract Study Group 1991, 'Risk Factors for Age-related Cortical,

Nuclear, and Posterior Subcapsular Cataracts', Am J of Epi,133, (6): p541-553.

Thompson J.R. 1989, 'The rote of registers in epidemiology: a disscussion pape/, J Royal

Soc Med,82: P151-152.

Thompson N. and Briscoe N. 1991,'Overview of Aboriginal Health in South Australia',

Abriginat and Tones Stæiit tstander Health Serþs No. 3:.Australian lnstitute of Heallh:

Thylefors B. 1985, 'Development of Trachoma Control Programs and the lnvolvement of

National Resources', Reviews ol lnlec Dis,7, (61:p774-776.

273 Relerences

Thylefors 8., Dawson C.R., Jones 8.R., West S.K., Taylor H.R. 1987, 'A simple system for the

assessment of trachoma and its complications', Bull World Health Organ,65 (4): p477-

483.

Thylefors B. 1987, 'A simplified methodology for the assessment of blindness and its main

causes', World Health Stat Q 40 (21: p129-141.

Thylefors B. 1990, 'Primary eye care and the design of the W.H.O. Programme for the

Prevention of Blindness', tnt OphthalmolClin, Winter, 30 (1): pl2-15.

Tielsch J.M., West K.P. Jr, Johson G.J., Tisazu J., Schwab L., Chirambo M.C., Taylor H.R.

1987, 'Trachoma grading: observer trials conducted in southern Malawi', Br J

Ophthalmol, MaY, 71 (5): P371 -374'

Tielsch J.M., West K.P., Katz J., Keyvan-Lariian Ê-, Tizazu T., Schwab L., Johnson G.,

Chirambo M., Taylor H.R. 1988, 'The epidemiology of trachoma in Southern Malawi',

Am J Trop Med Hyg,38: P393-399.

Tielsch J.M., Parver L.M., Shanker B. 1989, 'Time trends ¡n the incidence ol hospitalized

ocular trauma, Arch Ophthalmol, 1 07: p519-523.

Tozilto P. Kerr C. 1991, 'Conlempory issues in Aboriginal public Health' (Chapter 8). ln: IDe

Heatth of Aboriginat Austratia, eds Reid J., Trompf P., Centre for Cross-Cultural

Studies in Health and Medicine, University of Sydney' HBF (Aust).

Trachoma and Eye Health Report 1985. Nalio nal Trachoma and Eye Health Program Review

1985

Treharne J.D. 1985, 'The community epidemiology of trachoma', Review of lntecl Drs, Nov-

Dec : p760-7&4.

274 Ref erences

Turner V.M, West S.K., Munoz 8., Katala S.J., Taylor H.R., Halsey N., Mmbaga B.B.O. 1993'

'Risk Factors for Trichiasis in Women in Kongra, Tanzania: A Case-Control Stud/, lnf

J of Epidemiol,22 (2), P341-346.

Vaughan D. and Asbury T.1983, GeneratOpmhahnology- Lange Medical Publications.

Venkataswarmy V.G., Brilliant G.E. 1981 , Socialand Economic barriers to cataract surgery in

rural and south tndia. A preliminary report:Visual lmpairment and Blindness, p405-

508.

Verin p., Comte G., Nguyen Duy Tan. 'New Classification of Trachoma', Rev'lnt-Trach-Pathol'

Ocu t -T ro p- Su bt ro p- Sante - P ubliqu e, 1 -2: p 55-8 1 .

Vinger p.F. 1991,'sports eye injuries: a preventable disease', Ophthalmol SS: p2575'2577.

Viswalingain N.O., Wishart M.S., Woodland R.M. 1983, 'Adutt Chlamydial Ophthalmia

(Paratrachomal', Brit Med Bull,39 (2): p123-127.

Waite E.R. 1916, Irans Proc Roy Soc of SA, volxli: p429

Waite J.H., Beetham W.P. 1935, 'The visual mechanism in diabetes mellitus (a comparative

study of 2002 diabetics, and 457 non-diabetics for controls)', N Engl J Med,212:367-

79 : p429-438.

Watson D.S. 1 987, 'Apathy or antagonis m?' , Aust Family Physican, 1 6,5: p664-671 .

Watson P.G. 1969, 'Trachoma', Brit J Hosp Med, 2: pl650-1 656.

Webb S.G. 1990, 'Prehistoric eye disease (trachoma ?) in Australian Aborigines', Am J Phys

Anthropol, Jan;81 (1): p91-100.

Werner G.T., Sareen D.K. 1977, 'Trachoma in Pujab, a study of the prevalence and of mass

treatment', TroP Geogr Med, 29:. P35.

275 References

West S.K., Munoz 8.E., Newland S. et al. 1987, 'Lack of evidence for aspirin use and

prevention of cataracts', Arch Ophthalmol, 1 05: pl 229-1 231.

West S., Taylor H.R. 1988, 'Community-base intervention programs lor trachoma control', /nf

Ophthalmol, 1 2 (1 ): P19-23'

West S., Lynch M., Turner V., Munoza 8., Rapoza P., Mmbaga B.B.O., Taylor H.R. 1989'

,water availability and trachoma" Bull world Health organ,67 (1): p71-5.

West S.K., Taylor H.R. 1990, 'Reliability of photographs for grading trachoma in field studies',

Br J Ophthalmol,T4: P12'13.

West S.K., Congdon N.., Katala S., Mele L. 1991, 'Facial Cleanliness and Risk of Trachoma in

Families', Arch Ophthalmol, June, 109: p855€57.

West S.K., Munoz 8., Turner V.M., Mmbaga B.B.O., Taylor H.R. 1991, 'The Ëpidemiology of

Trachoma in Central Tanzania', lnt J of Ep¡demiol, 20 (4): pl 088-1 092.

White G.F. (ed) 1972, Drawers of watet: domestic watet use in East Aîñca., Chicago, University

of Chicago Press :P6-16.

Wilson M., Keyvan-Larijani E., Millan-Velasio F., Telsch J.M., Taylor H.R. 1987, 'The

Epidemiology of Trachoma in Chiapas (Mexico)', Rev-lnt-Trach-Pathol-Ocul'Trop'

Subtrop-Sante-Publique, &i p1 59-1 74.

Winkler P.G. 1963, 'A morbidity survey of trachoma and other communicable eye diseases in

the distric{ of Hebron, Jordon 1960', Bull World Health Organ,28:fi17.

Woolridge R.L., Grayslone J.T., Perrin E.8., Young C.Y., Cheng K.H., Chang LH. 1967'

'Natural History of Trachoma in Taiwan SchoolChildren', AmerJ Ophthalmology,May:

p1 31 3-1 320.

World Health Organ¡sation 1962, 'Expert Committee on Trachoma. Third report', World Health

Organ Tech Rep Sen : No. 234, W.H.O. Geneva..

276 Ref erences

World Health Organisation 1973, 'Prevention of Blindness', W H O Tech Rep Ser: No. 518,

W.H.O. Geneva

World Health Organisation 1979, Guidetines for Programs for the Prevention of Blindness,

W.H.O. Geneva.

World Health Organisation 1989, Primary Health Care Level Management of Trachoma.

W.H.O. Geneva.

Wyatt K. 1991, 'Aboriginal health research: Aboriginal and Torres Strait lslander community

involvement', PHA Conlerence Abstracts, Ausl J of Public Health,15,4;p342.

Yates P.C. 1963, 'Blindness in Australia', Med J Ausl, 1: p828- young E., Taylor H.R. 1984, 'tmmuns mechanisms in chlamydialeye infection: cellular imm¡ne

response in chronic and acute disease.', J lnfecl Die 150: p745'751.

Zigman S. 1973, 'Ocutar proetin alteration by near ultraviolet light', Exp Eye Res : @5&ü.

Zigman S., Datiles M., Torczynski E. 1979, 'sunlight and human cataracts', lnvest Ophthalmol

Vis Sci,18: P462-467.

277 Appendices

278 Tables of VisualAcuity Appendix 1

Table 1: Right vs left eye corrected visual acuity in the 0-19 age€roup

LCORRECTED 6/6 6/6 6t12 6t18 6/36 6/60 3/60 cF PL NPL RCORRECTED Uncorr. Corr. ToË

6/6 Oor. 756 0 0 0 0 00 00 0 756

6/6 tlcor. 1 6 0 0 0 00 00 8 6t12 0 I 0 0 0 00 00 0 1 6/1 I 0 0 0 0 0 00 00 0 0 6/36 0 0 0 0 0 00 00 0 0 6/60 0 0 0 0 0 00 00 0 0 3/60 0 0 0 0 0 00 00 0 0

CF 0 1 0 0 0 00 00 0 I PL 0 0 0 0 0 00 00 0 0 NPL 0 0 0 0 0 00 00 0 0

Total 757 I 0 0 0 0 0 0 0 1 766

Table 2: Right vs left eye corrected visual acuity in the 20- 29 age-group

LCORRECTED 6/6 6/6 6t12 6t18 6/36 6/60 3/60 cF PL NPL RCORRECTED tìær. Oor. Totd

6/6 tbr. 211 0 0 0 0 00 0 00 211 6/6 Oor. 0 3 0 0 0 00 0 00 3 6t12 0 0 0 0 0 00 0 00 0

6/1 I 0 0 2 1 0 00 0 00 3 6/36 0 0 0 0 0 00 0 00 0 6/60 0 0 0 0 0 00 0 00 0 3/60 0 0 0 0 0 00 0 00 0

CF 0 1 0 0 0 00 1 00 2 PL 0 0 0 0 0 00 0 1 .0 I

NPL 1 0 0 0 0 00 0 00 I

Total 212421000110221

279 Tables of VisualAcuity Appendix 1

Table 3: Right vs left eye conected visualacuity in the 30-39 age{roup

LCORRECTED 6/6 6/6 6t12 6t18 6/36 6/60 3/60 cF PL NPL RCORRECTED l.lor. Oor ToH

6/6 tìær. 125 00 0 0 0 0 0 00 125 6/6 Oor. 0 72 0 0 0 0 1 02 12

6112 0 13 1 0 0 0 0 00 5 6/1 8 0 11 't 0 0 0 0 10 4 6/36 0 00 0 0 0 0 0 00 0 6/60 0 00 0 0 0 0 0 00 0 3/60 0 00 0 0 0 0 0 00 0 CF 0 10 0 0 0 0 0 00 1 PL 0 00 0 0 0 0 0 00 0 NPL 0 00 0 0 0 0 0 00 0

Total 125 10 6 2 0 0 0 1 1 2 147

Table 4: Right vs left eye conected visual aanity ¡n the 40-49 age-group

LCORRECTED 6/6 6/6 6t12 6118 6/36 6/60 3/60 cF PL NPL RCORRECTED (lær. Oor. ToË

6/6 LÌrcor.gg0000000 0 0 99 6/6 Oor. 0 3 2 3 2 301 0 0 14

6112 0 2 3 f 0 1 0 00 0 7

6/1 I 0 0 1 2 1 1 0 0 0 0 5

6/36 0 2 0 1 0 0 0 0 0 0 3 6/60 0 0 0 0 0 0 0 0 0 0 0 3/60 0 0 0 0 0 0 0 0 0 0 0 CF 0 2 0 0 0 0 0 0 0 0 2

PL 0 0 0 0 1 0 0 0 0 0 1

NPL 0 1 0 0 0 0 0 0 0 0 1

Total 99 10 6 7 4 5 0 1 0 0 132

280 Tables of VisualAcuity Appendix 1

Table 5: Right vs lett eye conected visual acuity in the 50-59 age€roup

LCORRECTED 6/6 6/6 6112 6t18 6/36 6/60 3/60 cF PL NPL RCORRECTED [iær. Oor. ToH

6/6 t-frcor. 56 0 0 0 0 0 0 000 56

6/6 Oor. 0 10 3 1 0 0 0 0 20 16

6t12 0 3 I 2 1 1 0 0 1 0 16 6/1 I 0 5 3 0 0 2 0 0 0 1 11

6/36 0 0 0 0 1 0 0 0 0 0 1 6/60 0 0 0 2 3 0 0 0 0 0 5 3/60 0 0 0 0 0 0 0 0 0 0 0

CF 0 3 2 0 1 0 0 0 0 0 6

PL 0 0 0 0 0 0 0 1 0 0 I NPL 0 0 0 0 0 0 0 0 0 0 0

Total s6 21 16 5 6 3 0 f 3 1 112

Table 6: Right vs left eye conected visual acuity in the 60+ age€roup

LCORRECTED 6/6 6/6 6t12 6t18 6/36 6/60 3/60 cF PL NPL RCORRECTED tìær. Oo¡r Totd

6/6 LJlcor. 2300 0 0 0 0 0 0 0 23

6/6 Oor. 0 4 1 0 I 0 0 2 0 0 I

6t12 0 3 I 3 2 1 0 1 2 1 21

6/1 8 0 1 3 6 1 I 0 2 1 2 17

6/36 0 1 1 3 2 4 0 5 4 1 21

6/60 0 0 1 0 3 2 1 1 0 1 I

3/60 0 0 0 1 0 0 0 1 0 1 3

CF 0 2 1 1 2 0 0 7 0 2 15

PL 0 0 1 3 1 1 0 2 1 2 11

NPL 0 0 3 0 2 0 0 1 1 I

Total 23 11 19 17 14 9 1 22 9 11 136

281 Tables of VisualAcuity Appendix 1

Table 7: Right vs left eye corrected visual aotity for all agegroups combined

LCORRECTED

6/6 6/6 6t12 6118 6/36 6/60 3/60 cF PL NPL RCORRECTED tlær. Oor ToEl

6/6 l..hær.1270 0 0 0 000 0 00 127 0

6/6 Cor. 1 33 I 4 330 4 2 3 6l 6t12 0 10 22 7 330 1 3 1 50

6/1 8 0 7 10 10 240 2 2 3 40

6/36 0 3 1 4 340 5 4 1 25

6/60 0 0 1 2 621 1 0 1 14

3/60 0 0 0 1 0 0 0 1 0 3

CF 0 10 3 1 3 0 0 8 0 2 27

PL 0 0 1 3 2 1 0 3 2 2 14

NPL 1 1 3 0 2 0 0 1 1 1 10

Total 1272 64 49 32 24 17 1 26 14 15 1514

282 Estimates of Death Rates Appendix 2

The projected 1986 Census data was estimated approximately using 1986 census data. lt was assumed that the population was equally distributed in each age group, there would be no net migration which would affect the population distribution and the population list for 1990 could be used to estimate the size of the 0-4 age group. An estimate lor the total number of births could have been used, but there were several problems with birth and death registration data (Australian Bureau of Statistics 1988).

An estimate for the death rate was used (7.2110001, which was the rate for the Aboriginal population ¡n the Far North of South Australia in 1988, age-specific rates for lhe state were used to apportion deaths among the age-groups (Australian Bureau ol Statistics 1988). As the projected data was only a crude approximation, only deaths in the 45+ age groups were calculated and adjusted for.

The number of deaths in the four year period were calculated as

M F

50-59 14 16

60+ 24 12

The increase in size for each age group was calculated by 'moving" four years of one age category individuals into the next age-bracket until all age-groups were adjusted. Deaths were

adjusted for last of all.

283 Eye Examination Sheet Appendix 3

IDENTTY NO AGE M/F ABORIGINAL Y/N

OCCUPATION COUNTRY OF BIRTH

P.O,H. ACTION TAKEN

RL RL B. VISION E. TRACHOMA

I-INAIDED NO SIGN

CORRECTED (P.H. or Glasses) SIGN OFTRACHOMA TF TI

TS.. . C. I]ASIC EYE EXÄMINATION TT .... co .... NO ABNORMALITY: ...... NOTEXAMINED ABNORMALITY

EYELID: Intumcd margi ry't¡ichiasis INTRAOCULAR PRESSURE Neoplasm IOPinmmHg.... GLOBE: NOTEXAMINED ...... P h tl is ical/disorganized/absen t

CORNEA: Central comeal opacity OfflONAL EXAMINATION (<6i l8 either eye best conectcd) Pærygium (come¿I) . . DILATEDFUNDUS. LENS: No view of lens NOTEXAMINED .. Obvious opacity Aphakia..... Pscudophakia . REM/TRKS NOTEXAMINED

D. PREVIOUSEYESURCERY

NO EVIDENCE OF SURGERY

TYPE OF PREVIOUS SURGERY:

Eyelid . Cata¡act Glaucoma

Reúnal . Other ..

NOTASSESSED

284 Eye Examination Sheet Appendix 3

F vrs c. CURRENTACTTONNEEDED FOR THE R--;- PERSON: DISORDERS FOR FNCH EYE: MARK ONLY NO CURRENT ACTION NEEDED MÄRK ALL ONE WHICH PRINCtrÁL RL^PPLY DISORDER ACTIONNEEDED: P h rhjs ical, D iso rgariizcd, Eyelid surgery or Abscnt Globc . .... Rcf¡activeError,...... Caøractsurgery .. CåLaract Glaucoma t¡rirtrnent Uncorreaed Aphakia Spcct¿clcs orPseudophakia ...... Medic¿tion ...... Trachomatous Come¡l Opacity Other .. Oûrc¡ Cornc¿l Opacity ...... Antc¡ior UveitLs Glaucom¡ EXAMINER NO. Opdc Arophy

Vascul¡rRednopathy ...... REII{ARKS: Chorioretinitis Mâcub¡Dcgcnccaúon ...... Otìe¡ . .

NOTEXAMINED... T

PRINCIPAL MÁ.RK AI¿ C^USEM^RK UNDERLYING CAUS$: WHICI{ A.PPLY OIG ONLY

No lLsted underllng cause . Trauma CongenitaVNconatal Facor

Me¿sleslr'iu A Deficiency . Surgical Proccdure ...... DiabetcVllypencnsion .. . . Optional Facor

285 The Five Sign Trachoma Grading System Appendix 4

BETTER VISION FOR ALL

The Five Sign System for Trachoma and Age Specific Preval ence Rates

The five signs are: nl .r.F Follicular trachoma is scored whenever follicles are seen in the conjunctiva I f of the everted uPPer lid. o, when scars are seen in the coniunctiva or the TS ]lX:i:"J";rrrïlll: 'r '.o|'", (3) Trachoma intense is scored when no normal tarsal vessels are seen in the TI everted upper lid. ,0, Trachoma trichiasis is scored when one or more eyelash touches the TT eyeball. (s) dense enou gh to c o :;J:::J :il"ï'";',;:ï"fi ili:,"'Îl ;o;ïj."JI ,.l:;i:'*

An eve mav be scored for one or more of these signs Procódurei (1) Magnification in day light or torch light. (2) Only the upper lid conjunctiva is graded, excluding lateral and medial corners. (3) The presence of a sign in one or both eyes categorises the patient as Positive for that sign. (4) The right eye is examined and recorded first' (5) Tally is entered thus: ffi: s -lltf lll : s )+tW lll : t¡ etc Photographic examples

Normal upper lid conjunctiva. Note the normal vertical blood vessels. NORMAL Not scored for TF, TS or Tl

286 The Five Sign Trachoma Grading System Appendix 4

TRACHOMA FOLLIC[ES: Follicles are the pale areas about the size of a pinhead. These two eyes are scored for TF.

WASHING EYES AND IACE THREE TIMES A DAY MAY HELP YOU KEEP TRACHOMA AWAY

TRACHOMA INTENSE: No normal vertical blood vessels seen, Tl (in this case follicles (TF) obscure the vessels). Scored for TF and Tl

TRACHOMA TRICHIASIS: Eye lash or lashes touching the eyeball. Scored TT'

CORNEAL OPACITY: ls scored when opacity is CORNEAT OPACITY: ln this case of CO the opacity dense enough to prevent a view of the pupil' although not over the pupil is dense enough to ln this case the whole cornea is opaque. obscure it if it were so placed. Scored CO.

287 The Five Sign Trachoma Grading System Appendix 4

TRACHOMA SCARRING: The lines forming a network are scars. These two eyes are scored for TS.

TRACHOMA INTENSE: No norma I vertical blood TRACHOMA INTENSE: No normal vertical blood vessels seen, Tl (in this case due to papillae obscuring vessels seen. TS (in this case scars (TS) obscure the vessels). Scored for Tl only. vessels). Scored for TS and Tl.

REQUIREMENTS FOR DOMICILIARY HYCIENE

1. Water reticulated into the dwelling available at 100 litresþerson/day. 2. Means of heating water for body/clothes washing. 3. Showers to enable every person to shower at least once per day. One shower per 10 persons. 4. Washing machine for clothes and linen. 5. Sleeping surfaces that are elevated, separated NB: lf eye pulled out to and ventilated. touching the eyeball TT is scored 6. No more than 2 persons for each 3 x 3 metre square sleeping room. 7. Fly screens on all exterior doors and windows. 8. Cleanable surfaces inside dwellings. 9. Living space free of animals. 10. One toilet per 10 persons.

,,TRACHOMA GOES WHEN HOME HYGIENE rs cooD" CORNEAI OPACITY: From above (panus) also grades case as CO if dense enough, as in this case.

288 Ethical Approval/ Letters and Permits Appendix 5

ABORIGINAL HEALTH ORGANISATION OF S.A.

Head Oll¡ce: 62 Beulah Fload, Norwood, S,A 5067 P O Box 285. Norwood, S.A. 5067 Telephone: reler:87'165-f 333 7300 Fax: 363 1514 tn rcpty pteasa Cuote: 14/95/04

Dr Nigel Stocks Dept of Prinrary HeaJ-tà Care Flinders Medical Centre BEDFORD PARK SA 5042

Dear Dr. Stocks,

RE: PROPOSED VISUAL II,IPAIRMEIÍI STUÐY IN ABORIGIòIAL COMMLJNITIES I apologise for this late response. As t}le Conmittee had indicated to you when you presented your proposaJ-, the Study is supported by the Aboriginal tþalth Researcà Conmittee. the Conmittee EeLt that thre aqtion

Best Wishes

},1. HAT{PIþN CHAIRPERSON ABORIGINAL HEALTH RESEARCH COMMITTEE

CDB: MH : LD

22nd August, 1989

289 Ethical Approval / Letters and Permits Appendix 5

Fl inders Med ical Centre BEDFORD PARK SOI'TH AUSTRALIA 5042 TELEPHONE 27199t1

2? September 1ff39

MEMORANDUM

TO: Dr. N. Stocks. Þsrtrûsot of Priraæy Health Care

IROM: Dr. A. Vedig, Chairman, Coornille€ on Clinical Investigations

TOPIC: P.e¡earch Applicatior ?l /89

Your atteûtion is dravn to lhe folloving extract from lhe Mi¡ues of the Comroinee's meeting held on 2J SepterûÞer lS9.

1899 I Reseæch Aoolicatioo ?lÆ9 - Dr. N. Stocks A population based prercleoce sr¡rræy of visual imgairmenl in the n¡rsl Aborigi¡al population of South Äwralia

Lener frûm Dr. Slocks outlining the i¡formation this Comfûitlee requested ¡¡as received and the agplication aeproved.

Ä. Vedie lntensnæ Cme Ilnit. Ertension {?{

290 Ethical Approval/ Letters and Permits Appendix 5

AI{AI{GU PITJAf{fJAÏJAßA l{o s0utlt AUSlSAUA P|ÍJAÍ{IJAÍJARA r.A¡0 RlGtfTS rgü PERMIT TO Ef{TER OTTO AflO REMAIT Of{^CT, ASORICII¡AT I¡f{O Punu¡nt Io 0ivi¡ioo ll ÒçY=E6= > s1 6txEß^t c0¡0tTr0fs 0f rsSUt: 9c.= o I lh'r do.r nol .urhonr. rha âñrry ol ¡ parson ro ¿ dw.ll'.9 or |vrñC .ra¡ a ocça Þ.rmrl ol ! c ¡ñ0 Õccuo,ôd by o, Þ.1ory'ñg lo .ñ Aboû0rñ.1 wrthoul lh! conr.nl ol lh. ow^ar 9 V c- ) 2 Ihrr pêrn't r¡ sùbt.cl ro tñ. sp¿cr.l coñdrr'o^r (rl ¡ñy¡ ¡ar oul b.low ! I So.c,.l cond'tions o PURPOSIS Of EilTNY h aõ ã É õ I ?.: 4 lrsuêo'lh'rp6rñrrdo..norrmÞlyrhrr¡or¡ccol'^r.nrroñtovr!'rth.Abo¡rgrn¡¡ 0AfES: from iåñó sÞ.crlrd h.r b.cñ srñ.d úoo^ lh. l¡¡d¡koñ¡l owñart conc.ñ.d Th? = p.¡ñrrraê ¡r ..rpoñ3¡bl. ro rh¡r Abo¡0rñ.1 coññuñrl¡.a ..a rnlorñad ol hrt añrqr. s>:: a 55 ued al r9 ¡nlañlrOñ lO vrsrl f, th'¡ v!lrd p.dorm lhc ¡a!ocr.l.d 5 Þ.mrr 't oñly ro .ñ¿bl. rh. Þ..ñ'(r. to duir.a w'rh ì'r o. hrr tr.lad Þurpo¡. Srgned by 6 Thr! oarñ,r murr ba c¡r.r.d ar all rrñ.s wh'1. rh. hold.. r¡ on lha Abong¡n¡l l¡nd .ñd o.oduc.d lor ¡^so.cùon o^ d.ñ.ñû 3< -: fh. p..ñrtr.á ñ!!r comply w'th .ñy l.wr curr.ñily lorca coñcctñ'ñ0 th. t.là u o ó.: 7 'n ând ætr.t!ro^ ol lrauo, o^ rh. abo.ror^¡l l.ñd Signature I Th,¡ Þ.mrr ñ.y b...vo¡.d ¡l .ny r,ñ. 9 lñ rñ. .vêñr ol ..voc.lroñ ol . o.rh'r o, ¡r! c..!rng lo Þ. vrl'd lot aôY r..to¡ lha p..ñr(a. rh.¡l rññ.d'.r.lv 1..v. rñ. P'(t¡.rt¡ltrr. l¡ñd '.aahold

291 Ethical Approval/ Letters and Permits Appendix 5

The lJniversity of Adelaide BOX 49S, O P.O.. ADEL-A.IDE, SOUTH AUSTRALI.A 5OOI

Rel:LDZ D.2û3n5 Enquiries: Ms. L. D. Zeiu Sscta¡Y, Comminee on thc Ethics of Human Experimentarion Phone: (08) 228 5184

THE REGISTRY 21 September 1 989

Dr. N. Stocks Department of Primary Health Care School of Medicine Flinders University of South Australia Bedford Park SA 5042

Dear Dr. Stocks, Hl3TlSg A Population Based Prevalence Survey of Visual lmpairment in the Rural Aboriginal Population of South Australia

Thank you for your letter of 13 September 1989.

I am pleased to inform you that the Chairman on behalf of the Committee for the Ethics of Human Experimentation has approved the above project.

please nole that any chan_oe to the project which may affect its elhical aspects will invalidate the project's approval. ln such cases an amended protocol must be submitted to the Committee for further approval.

Where possible, sublects taking part in the study should be given a copy of the lnformat¡on Sheet and the s¡gned Consent Form to reta¡n.

Project approvals are current for one year. The expiry date for your project will be 30 September 1990. Applications for renewal must be accompanied by a brief reporl on the pro¡ect's progress and any ethical issues which may have arisen. lf the proiect is completed before that time, a brief report on ¡ts outc¡me and the ethical considerations should be made to the Comm¡ttee.

Itake lhis opporlun¡ty to wish you well in your research. Yours sincerely,

F.J. O'NEILL Registrar

Telephonc: (08) 228 51ll FÀX: (03) :lJ 0464 Tclcr: UNIVAD \i',i

292 (f) (o o) x ô¡ Þ c (D o- o-

+ -t

vrlvu lsñv Hlnos N soñvr vuvalv^NVvoN oñv vuvf lvlnv9Nn^ vBvrrvrlNvf 1td

¡

I I À- r¡,!.ç,-l'.t CL c (ú t

ÑB]HlUON ^uolrutll Community Newsletter: Article on Trachoma Appendix 7

Í.^cho(\q [^,^. \s ô \-\<- 17{s /

-lLI \s 4. re-o. J v/ {\\<-s d.

\ \^ q.\-i A\"+ "{ P.O P L e-\-e- L tro. cl- c¡ v-\ <:- l-t c-<>'- -t ,/ "u t\, '^J. â1tè- \w\e- \- a-<-,^* t o- t\ ^ rÌ lr¡.*-tn S c..S\-,.-s l\-..¿- ê .v 4 '/ -1\ Vr,e_ce_ rS nre-Å'. (-in€-.

o N I tJc.s\r {^.-- ..^ A l\^ e.- u"'7 {".*5 lo Or¿ u¡ \ "^\ \, \ao. cFrorv\a

294 Community Newsletter: Art¡cle on Trachoma Appendix 7

a-r zt A.l.\^.,1- \ o.^ N5o^ aerf ft!"^å- 3*-"^ c^q^b ¿( a- !.* oçoÁ k-* o^^ì_J- 7 K ¡. , N/ !,, P. I , ¿\r\1.\ ' V ^*

'\v\-\ D.t¡o . .- ß. "-..i

i'LqqI GtL.- int¡ .J 1 _ K'ar r\ c'¡] ''-i-jct va

295